JP3502468B2 - Distributed monitoring equipment - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distributed monitoring apparatus used for security, management, maintenance, etc. using images in buildings, public facilities, shops, parking lots and the like.

[0002]

2. Description of the Related Art In recent years, the necessity of an automatic monitoring device for facilities has been increasing year by year due to the enlargement of facilities and soaring labor costs. As an automatic monitoring device for the equipment currently put into practical use, for example, an infrared sensor for detecting a person, a loop coil responsive to the magnetism of an automobile, and the like can be given. However, these sensors require extensive work to install,
There are problems that maintenance is troublesome, and that it is difficult for a supervisor to judge the situation only by sensor output when an abnormality is detected.

In order to cope with these problems, a monitoring device has been proposed which uses a television camera or the like which inputs a moving image which is easy to install and maintain and which is easy for humans to understand. As an example thereof, there is an image monitoring method described in Japanese Patent Publication No. 3-37357 which combines an infrared camera and a visible light camera, and each pixel of an image at a certain time is used as a reference image or a past image corresponding to the background. Comparison is made, and when the change is large, it is determined to be abnormal.

Further, as an example of a monitoring device for monitoring a wide area that cannot be detected by one camera, an image monitoring device for extracting a change area and a distributed network system are used in Japanese Patent Laid-Open No. 4-261289. Image monitoring system, and a distributed image recognition system described in Japanese Patent Application Laid-Open No. 2-82375.

On the other hand, in recent years, in the field of computer vision research, research aiming at more accurate image recognition by integrating images from a plurality of observation points and processing results processed by a plurality of algorithms has been popular. Is becoming.
As an overview article, for example, "Image Understanding by Distributed Cooperative Processing", Takashi Matsuyama, Measurement and Control, Volume 31, pp.1149-1154.
(1992).

[0006]

However, conventional monitoring devices, including the monitoring devices described in the above references, have the following problems.
It is not a practical automatic monitoring device because it is impossible to track when multiple people or vehicles move around in a wide area due to changes in the environment such as lighting conditions. The main problems in constructing a practical automatic monitoring device will be described below.

First, even if a conventional monitoring device targets a plurality of areas, since the information obtained in each monitoring area is independently used, an abnormality is detected in each monitoring area. If that fails, I couldn't fix it. For example, in the image monitoring system described in JP-A-4-261289, when a moving object is detected in an area monitored by an image monitoring apparatus, the moving object is tracked and other areas are monitored. Although the function to put the image monitoring device in the standby state is described in the embodiment, since the moving object is detected as a mere brightness change region, it is difficult to predict the movement and the moving object is stable. Since it becomes possible to trace the moving object only after it is detected by, there is a problem that it does not work well if the moving object itself is unstable.

Secondly, the conventional monitoring device detects an animal body from an image input in a relatively short time, and therefore, when the animal body temporarily passes by or is hidden by a stationary object. It failed to detect the animal body. Then, even if the passing of each other is over and the moving body can be detected again,
It was impossible to deal with the moving body before and after passing each other. In the field where the monitoring device is actually used, events such as passing each other frequently occur, and there are also many requests for information on the route of each moving animal body, which compensates for the temporary failure of detecting the moving body. I could not maintain time consistency.

Thirdly, in the area monitored by the monitoring device, the lighting conditions are often not constant, especially in the outdoors, so that the monitoring is impossible depending on the weather conditions and the time zone. In addition, it requires a lot of trouble and time to make adjustments such as installation. For example, the image monitoring method disclosed in Japanese Examined Patent Publication No. 3-37357 aims to provide a more reliable image monitoring method by combining an infrared camera and a visible light camera. Since it is assumed that the detection is stable, it is difficult to monitor at night, for example. Further, since there is no function for adjusting the camera using the result of the detection of the moving object, the camera itself is not properly adjusted and the input image itself is inferior. In addition, JP-A-2-823
The distributed image recognition system disclosed in Japanese Patent Publication No. 75 is composed of a plurality of image recognition devices and a centralized management device, and the centralized management device sends tuning parameters to the image recognition device based on the processed image from the image recognition device. The processed image output from the image recognition device is only an image from which contour lines are emphasized and noise is removed, has no relation to higher-level processing, and the use of information on moving region processing and moving object recognition and tuning parameters. It lacked the concreteness of the creation process.

Fourth, the module design in which the system is divided into subsystems is insufficient, and the interface between the subsystems is not abstracted, so that it is difficult to design and change the system. Conventional monitoring systems were often designed and manufactured only once, and such a problem did not become apparent.However, due to the growing demand and expansion of demand for inexpensive monitoring devices, There is a demand for a highly flexible system that requires minimal changes to the overall system even if the number and type of changes and unexpected additions of image input devices and image processing devices are made.

The present invention solves the above-mentioned conventional problems. First, by integrating the results of the moving image processing of a plurality of moving image inputting means, a distributed type monitoring device with a few moving object detection errors. The purpose is to provide.

Secondly, by integrating past animal body information, there is little error in the detection of the moving body, and even if the moving body cannot be detected temporarily from the moving image, it is consistent in time. It is an object of the present invention to provide a decentralized monitoring device capable of detecting a moving object and tracking the moving object even if the moving object is complicated.

Thirdly, by abstracting information among the moving image input means, the moving image processing means, and the moving object detecting means, it is possible to easily deal with design changes and expansions, and it is highly flexible and reliable. It is an object of the present invention to provide a distributed monitoring device with high efficiency.

Fourthly, the state of the moving image input means and the state of the moving image processing are changed by using the result of the moving image processing and the result of the moving object detection, thereby saving labor in the adjustment at the time of installation. At the same time, it is an object of the present invention to provide a distributed monitoring device that is highly flexible and has a small error in detecting a moving object by following changes in the environment.

[0015]

In order to achieve the first to third objects, the distributed monitoring apparatus of the present invention has the following configuration. The distributed monitoring device of the present invention is
First, a plurality of moving image input means, and moving image processing means for outputting moving area information including a moving area corresponding to a moving object and its feature amount from each moving image output from each moving image input means, The moving object information output means outputs moving object information from moving area information, and moving object information output means outputs moving object information storage means.

Secondly, the distributed monitoring apparatus of the present invention is connected to each of the M moving image input means and each moving image input means, and the moving area corresponding to the moving object is output from each moving image. And M moving image processing means for outputting moving area information including the feature amount, and M moving objects connected to each moving image processing means and outputting moving object information from the output moving area information. N detecting means, communication means for communicating between M moving body detecting means, and N moving body information output from the moving body detecting means (where N is an integer of M or less)
And animal body information storage means.

Thirdly, in the distributed monitoring apparatus according to the present invention, in addition to the first or second configuration, the moving image processing means is arranged so that the spatial position of the object corresponding to the area from the area in the image. It has an area position calculating means for calculating.

Fourthly, in the distributed monitoring apparatus of the present invention, in addition to the third configuration, the area position calculating means is arranged on an image corresponding to a point at which the object imaged in the image is in contact with a plane. It has a plane contact point coordinate calculating means for calculating coordinates and a space coordinate calculating means for converting the coordinates on the image, which is the output of the plane contact point coordinate calculating means, into the space coordinates of the area where the image is picked up.

Fifth, in the distributed monitoring apparatus of the present invention, in addition to the first or second configuration, the moving object detecting means estimates the current moving object information of the moving object detected in the past. The moving object information that is detected in the past from the moving object information estimating means, the moving object information estimated by the moving object information estimating means, and the moving area information output from the moving image processing means corresponds to which moving area the current moving object is. And an animal body motion area associating means for estimating whether or not to perform.

Sixthly, in the distributed monitoring apparatus according to the present invention, in addition to the third or fourth configuration, the moving body detecting means estimates the current moving body information of the moving body detected in the past. The moving object information that is detected in the past from the moving object information estimating means, the moving object information estimated by the moving object information estimating means, and the moving area information output from the moving image processing means corresponds to which moving area the current moving object is. And an animal body motion area associating means for estimating whether or not to perform.

Seventhly, the distributed monitoring apparatus of the present invention has, in addition to the sixth configuration, a moving object position estimating unit for estimating the current position of the moving object detected in the past by the moving object information estimating unit. is doing.

Eighthly, in the distributed monitoring apparatus of the present invention, in addition to any one of the fifth to seventh configurations, the moving image processing means changes the color of the object corresponding to the area from the area in the image. Having a region color calculation means for calculating, the moving object information estimation means,
It has an animal body color estimation means for estimating the color of the animal body detected in the past.

Ninth, in the distributed monitoring apparatus of the present invention, in addition to any one of the fifth to seventh configurations, the moving image processing means changes the shape of the object corresponding to the area from the area in the image. The moving body area association unit includes a moving body shape estimating unit that estimates the shape of the moving body detected in the past.

In the distributed monitoring apparatus of the present invention, tenthly, in addition to any one of the fifth to seventh configurations, the moving image processing means changes the texture of the object corresponding to the area from the area in the image. The moving object information estimating means has a region texture calculating means for calculating, and the moving body information estimating means has a moving body texture estimating means for estimating the texture of the moving body detected in the past.

Eleventhly, in the distributed monitoring apparatus of the present invention, in addition to the configuration of any one of the seventh to tenth aspects, the moving object associating means of the moving body is arranged in the area imaged by the moving image inputting means. Hiding object position storage means for storing the position of a certain hiding object,
The moving image input means position storing means for storing the position of the moving image input means and the current position of the moving object estimated by the moving object position estimating means are moving image input means by the concealed object stored in the concealed object position storing means. And a concealment judging means for judging whether or not it is concealed.

In a twelfth aspect of the distributed monitoring apparatus of the present invention, in addition to the configuration of any one of the seventh to tenth aspects, in the past when the moving body moving area associating means is estimated by the moving body position estimating means. It has a moving object hiding determining unit that determines whether the moving object is hidden by another moving object detected in the past from the current position of the detected moving object.

Thirteenth, in the distributed monitoring apparatus of the present invention, in addition to any one of the seventh to tenth configurations, a point on a certain space corresponds to which pixel of the moving image of each moving image processing means. It has an image coordinate calculating means for calculating whether to perform.

In the distributed monitoring apparatus of the present invention, in a fourteenth aspect, in addition to any one of the seventh to tenth configurations, the moving object associating means for moving objects stores the position of each moving image input means. Image input means position storage means and distance threshold value storage means for storing distance threshold values corresponding to each moving image input means are provided.

In order to achieve the above first to fourth objects, the distributed monitoring apparatus of the present invention has the following configuration. Fifteenthly, the distributed monitoring apparatus of the present invention is, in addition to any one of the first to fourteenth configurations, a moving image input adjusting means for changing a moving image input state of the moving image input means, and a moving image processing. And a moving image processing evaluation means for calculating an evaluation value of the calculation result of the moving area information of the means.

Sixteenth, the distributed monitoring apparatus of the present invention comprises, in addition to any one of the first to fourteenth configurations, a moving image input adjusting means for changing the state of moving image input of the moving image input means, And an animal body detection and evaluation means for calculating an evaluation value of the animal body information output from the animal body detection means.

Seventeenth, the distributed monitoring apparatus of the present invention is, in addition to any one of the first to fourteenth configurations, a moving image processing adjusting means for changing the calculation state of the moving area information of the moving image processing means. And a moving object detection evaluation unit that calculates an evaluation value of the detection result of the moving object that is the output of the moving object detection unit.

In the eighteenth aspect, the distributed monitoring apparatus of the present invention has, in addition to the seventeenth configuration, a moving image input adjusting means for changing the moving image input state of the moving image input means.

In a nineteenth aspect of the distributed monitoring apparatus of the present invention, in addition to the first to fourteenth configurations, moving image processing adjusting means for changing the state of calculation of moving area information of the moving image processing means,
And a moving image processing evaluation means for calculating an evaluation value of the calculation result of the moving area information of the moving image processing means.

[0034]

With the above first and second configurations, the distributed monitoring apparatus of the present invention is, firstly, the result obtained by processing the moving images obtained by the plurality of moving image inputting means by the moving image processing means, By integrating in the body detecting means, even if the result of processing the moving image input by a certain moving image input means by the moving image processing means is insufficient, the moving image input by another moving image input means is used. By using the obtained information, it is possible to perform monitoring with less error in detecting a moving object. Secondly, by integrating the past moving body information stored in the moving body information storage means by the moving body detection means, it is possible to supplement it even if the detection of the moving body fails temporarily. It is possible to perform monitoring capable of tracking the moving body even for complicated movement of the moving body. Thirdly, since the outputs of the moving image input means, the moving image processing means, and the moving object detecting means are abstracted as a moving image, a moving region, and a moving object, respectively, there are a plurality of types for each of the three means. Algorithms and hardware may be used in combination, and even if the algorithms and hardware used in each of the above-mentioned three means are changed, the influence on other means can be minimized and flexibility is increased. A highly reliable monitoring device can be provided.

With the above-mentioned third structure, the distributed monitoring apparatus of the present invention has a fourth structure in addition to all or part of the respective functions.
In addition, by integrating the position information of each moving area calculated by the area position calculating means, it is possible to perform highly reliable monitoring using the position information.

With the above-described fourth structure, the distributed monitoring apparatus of the present invention has the fifth function in addition to all or part of the respective operations.
In addition, by using the plane contact point coordinate calculation means and the spatial coordinate calculation means, the position of the moving body corresponding to each moving region can be easily obtained, and the position information is used by integrating the obtained position information. The monitoring can be performed with high reliability.

With the fifth configuration, the distributed monitoring apparatus according to the present invention has a sixth or sixth embodiment in addition to all or part of the above-mentioned operations.
In addition, the moving object information estimating means estimates the moving object information, and the moving object moving area associating means associates the estimated moving object information with the moving area information output by the moving image processing means at the present time. With this, it is possible to track an animal body and perform highly reliable monitoring.

With the sixth configuration, the distributed monitoring apparatus according to the present invention has the seventh or seventh feature in addition to all or part of the respective actions.
In addition, the moving object information estimating means estimates the moving object information, and the moving object moving area associating means associates the estimated moving object information with the moving area information output by the moving image processing means at the present time. Then, the moving object is associated with the moving object, the current position of the moving object is estimated from the result using the position information of the corresponding moving object, and the moving object is tracked. It is possible to perform highly accurate monitoring.

With the seventh configuration, the distributed monitoring apparatus according to the present invention has the eighth function in addition to all or part of the respective operations.
In addition, the current position of the moving object previously detected by the moving object information estimating unit is estimated, and the estimated current position is used for associating the estimated moving object information with the moving region information in the moving object moving region associating unit. As a result, it is possible to associate the moving body with the moving body with higher reliability, and as a result, perform highly reliable monitoring.

According to the eighth structure, the distributed monitoring apparatus according to the present invention has the ninth function in addition to all or part of the respective functions.
In, the color of the moving area is calculated by the area color calculating means, the color of the moving body is further estimated by the moving body color estimating means, and the estimated moving body information and moving area information in the moving body moving area associating means By using the similarity between the color of the moving area and the color of the moving object for the matching, the moving object can be matched with the moving object with higher reliability, and as a result, highly reliable monitoring can be performed.

According to the ninth configuration, the distributed monitoring apparatus of the present invention has the first and second functions in addition to all or part of each of the operations.
0, the shape of the moving area is calculated by the area shape calculating means, the shape of the moving body is further estimated by the moving body shape estimating means, and the moving body information and moving area information estimated by the moving body moving area associating means are obtained. By using the similarity between the shape of the moving area and the shape of the moving object for the matching, the moving object and the moving area can be associated with higher reliability, and as a result, highly reliable monitoring can be performed. .

According to the tenth structure, the distributed monitoring apparatus of the present invention, in addition to all or part of each of the above-mentioned operations, eleventh, calculates the texture of the moving area by the area texture calculating means, and further calculates the moving object. The texture of the moving body is estimated by the texture estimating means, and the estimated moving body information and moving area information in the moving body moving area associating means are associated with each other.
By using the similarity between the texture of the moving area and the texture of the moving object, the moving object can be associated with the moving object with higher reliability, and as a result, highly reliable monitoring can be performed.

According to the eleventh configuration, in the distributed monitoring apparatus of the present invention, in addition to all or part of each of the operations, twelfth, the estimated current position of the moving body estimated by the moving body position estimating means is displayed. The concealment determining means determines whether or not it is concealed by the concealing object at the position stored in the concealing object position storing means, using the position of the moving image inputting means stored in the moving image inputting means position storing means. When it is determined that the moving object is hidden by the hidden object in the image of the moving image input means, it is determined that the corresponding moving area does not exist and the moving object is not associated with the moving object. Since it is possible to prevent an error in associating with a moving area of, it is possible to perform highly reliable monitoring.

According to the twelfth structure, the distributed monitoring apparatus according to the present invention has, in addition to all or part of each of the above-mentioned operations, thirteenth, the moving object detected in the past estimated by the moving object position estimating means. From the estimated current position, the moving object hiding determining unit determines whether the moving object is hidden by another previously detected moving object, and if the moving object is hidden by another moving object, take action. Since it is determined that there is no moving area to be moved and the moving area is not associated with the moving area, an error in associating with the moving area of the moving object can be prevented, and thus highly reliable monitoring can be performed.

According to the thirteenth configuration, the distributed monitoring apparatus according to the present invention is, in addition to all or part of each of the operations, fourteenth, in the fourteenth aspect, the moving object detected in the past estimated by the moving object position estimating means. The coordinate position in the image coordinate system of each moving image input by each moving image input unit corresponding to the estimated current position is obtained by the image coordinate calculating unit, and the moving object detected in the past by the moving object moving region associating unit is It is determined whether or not each moving image is captured, and when it is not determined that the moving object detected in the past is captured in each of the moving images, it is determined that the corresponding moving region does not exist. By not associating the moving area with the moving area, it is possible to prevent an error in associating with the moving area of the moving object, and to perform highly reliable monitoring.

According to the fourteenth structure, the distributed monitoring apparatus according to the present invention has, in addition to all or part of each of the above-mentioned actions, fifteenth, the moving object detected in the past estimated by the moving object position estimating means. The distance between the estimated current position and the position of each moving image processing means stored in the moving image input means position storage means is obtained, and this distance corresponds to each moving image input means stored in the distance threshold storage means. If it is larger than the distance threshold, it is determined that the moving object previously detected by the moving object associating unit is far from the camera and the moving area is not calculated, and the moving object detected in the past is associated with the moving area. By not performing this, it is possible to prevent an error in associating with the moving region of the moving object, and to perform highly reliable monitoring.

According to the fifteenth configuration, the distributed monitoring apparatus of the present invention evaluates the result of the processing of each moving image by the moving image processing evaluation means in addition to all or part of the above operations. If, as a result of this evaluation, it is determined that the moving image input state should be changed, the moving image input adjusting means changes the moving image input state to improve the moving image input state. Therefore, the monitoring can be performed with high reliability even in an environment where the lighting conditions and the like change, and the labor of adjustment such as installation can be reduced.

According to the sixteenth structure, the distributed monitoring apparatus of the present invention, in addition to all or part of each of the above-mentioned operations, seventeenth, evaluates the result of the detection of the moving body by the moving body detection evaluating means, As a result of this evaluation, when it is determined that the moving image input state should be changed, the moving image input adjusting unit changes the moving image input state to maintain the moving image input state in a better state. Therefore, it is possible to perform highly reliable monitoring even in an environment in which lighting conditions and the like change, and it is possible to reduce the time and effort for adjustment such as installation.

With the seventeenth structure, the distributed monitoring apparatus according to the present invention, in addition to all or part of each of the above-mentioned operations, eighteenth, evaluates the detection result of the moving body by the moving body detection evaluation means, As a result of this evaluation, when it is determined that the moving image processing state should be changed, the moving image processing adjustment unit changes the moving image processing state to keep the moving image processing state in a better state. Therefore, it is possible to perform highly reliable monitoring even in an environment in which lighting conditions and the like change, and it is possible to reduce the time and effort for adjustment such as installation.

With the eighteenth configuration, the distributed monitoring apparatus according to the present invention, in addition to all or part of each of the above-mentioned operations, nineteenth, evaluates the detection result of the moving body by the moving body detection evaluation means, As a result of this evaluation, when it is determined that the moving image processing state should be changed, the moving image processing adjusting unit should change the moving image processing state and the moving image input state should be changed. If it is determined that the moving image input adjusting means changes the moving image input state, the moving image input and the moving image processing state can be maintained in a better state, and the lighting condition is changed under the environment. However, it is possible to perform highly reliable monitoring and reduce the time and effort required for adjustment at the time of installation.

According to the nineteenth structure, the distributed monitoring apparatus according to the present invention evaluates the result of the processing of each moving image by the moving image processing evaluation means in addition to the whole or a part of each operation. If, as a result of this evaluation, it is determined that the state of the moving image processing should be changed, the state of the moving image processing is changed by the moving image processing adjustment means to improve the state of the moving image processing. Therefore, the monitoring can be performed with high reliability even in an environment where the lighting conditions and the like change, and the labor of adjustment such as installation can be reduced.

[0052]

【Example】

(First Embodiment) A first embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing the configuration of a distributed monitoring apparatus according to the first embodiment of the present invention. Figure 1
In 1 and 7, video images of the surveillance area are captured, and YUV
Moving image input means for outputting as a color digital video signal of the format, 2 and 8 are moving image processing means for processing the moving images captured by the moving image input means 1 and 7 and outputting moving area information, 3 is a moving image. The present moving body is detected from the moving area information output from the image processing means 2 and 8 and the past moving body information stored in the moving body information storage means 4, and the moving body information is calculated and output to the outside. Animal body detection means, 4 is an animal body information storage means for storing animal body information of an animal body detected by the animal body detection means, and 5 is animal body information which is the result of the animal body detection by the animal body detection means 3, and an animal. Animal body detection and evaluation means for evaluating the moving body detection in the moving body detection means 3 from the past moving body information stored in the body information storage means 4, 6 is the moving image processing means 2 and 8 and the moving body detection means 3 And the moving object detection and evaluation means 5 A local area network for (LAN).

Hereinafter, the moving image input means 1 and 7 and the moving image processing means 2 and 8 will be described first, and then the moving body detection means 3, the moving body information storage means 4, and the moving body detection evaluation means 5 will be summarized. Explain.

(Moving Image Input Means) First, the moving image input means 1 of FIG. 1 will be described with reference to FIG. Since the moving image input means 7 has the same configuration and operation as the moving image input means 1, duplicated description will be omitted.

FIG. 2 is a diagram showing the structure of the moving image input means 1. In FIG. 2, reference numeral 11 denotes a television camera having a function of adjusting an aperture from the outside, which captures a moving image and outputs it as a YUV format color analog image signal, and 12 denotes a YUV format color analog image output from the television camera 11. It is an A / D converter that converts a signal into a YUV color digital image signal. Reference numeral 13 is a moving image input adjusting means for changing the state of the moving image input, and an aperture adjusting device 14 for adjusting the aperture of the television camera 11 and an amount of aperture adjustment of the aperture adjusting device 14 for detecting the moving object in FIG. And an aperture adjustment amount calculation circuit 15 that calculates the moving object detection evaluation value that is the output of the. Reference numeral 21 is a communication means for sending the output of the A / D converter 12 to the moving image processing means 2 of FIG. 1, and 33 is the moving object detection and evaluation means 5 of FIG. 1 transmitted through the moving image processing means 2.
Is a communication means for narrowing down the moving object detection evaluation value which is the output of the above and sending it to the adjustment amount calculation circuit 15.

Next, the operation of the moving image input means 1 will be described. In the present embodiment, the television camera 11 of the moving image input means 1 and the television camera of the moving image input means 7 image the same surveillance area from different directions. TV camera 11
The color analog video signal imaged in (1) is converted into a digital signal by the A / D converter 12, and is sent to the moving image processing means 2 via the communication means 21. On the other hand, from the moving image processing means 2, the moving object detection evaluation value which is the output of the moving object detection evaluation means 5 is input to the aperture adjustment amount calculation circuit 15 via the communication means 33. In the present embodiment, the evaluation value for detecting a moving object is
As a result of the detection of the animal body, it is a value indicating the degree of excessive detection of the animal body, the underdetection of the animal body, or the detection of a reasonable number.

Aperture adjustment amount calculation circuit 15 and aperture adjustment device 1
4 calculates the aperture adjustment amount of the television camera 11 and adjusts the aperture of the television camera 11 according to the algorithm classified according to each case as described below.

1) The moving body detection evaluation value indicates that the moving body is excessively or under-detected, and when the aperture adjustment has not been performed in the near past, an adjustment amount is randomly generated, The television camera 1 through the aperture adjusting device 14.
Adjust the aperture of 1.

2) If the moving object detection evaluation value indicates that the moving object is excessively or under-detected, and if the aperture adjustment has been performed in the near past, the moving object detection evaluation value and the past moving object detection evaluation value If the moving object detection evaluation value is improved or does not change from the past due to the aperture adjustment performed in the near past, the same aperture adjustment amount as in the past is output, and if it becomes worse, that is, When the absolute value of the moving object detection evaluation value is large, the aperture adjustment amount that is the opposite of the past positive / negative is output, and the aperture of the television camera 11 is adjusted via the aperture adjustment device 14.

3) If the moving object detection evaluation value indicates that an appropriate number of moving objects have been detected, the aperture adjustment amount calculation circuit 15 outputs 0 as the aperture adjustment amount, and thus the aperture adjustment is performed. The device 14 does not perform aperture adjustment.

As described above, the moving image inputting means 1 in the present embodiment captures an image of the surveillance area and outputs the result as a color digital video signal, and at the same time, the aperture of the television camera is adjusted using the moving object detection evaluation value. By setting the optimal state, it is possible to output a moving image in which the moving object is appropriately detected even if the illumination condition of the monitoring area changes.

(Moving Image Processing Means) Next, the moving image processing means 2 of FIG. 1 will be described with reference to FIG. Since the moving image processing means 8 has the same configuration and operation as the moving image processing means 2, duplicate description will be omitted.

FIG. 3 is a diagram showing the structure of the moving image processing means 2. 3, 22 is a frame buffer for storing a frame image of one frame of the color digital video signal output from the A / D converter 12 of FIG. 2 which is input via the communication means 21, and 23 is the frame buffer 22. A frame buffer that temporarily stores the frame image stored in the frame buffer 22 while the next frame image is stored in the frame buffer 22, and 24 represents a brightness value of each pixel of the frame images stored in the frame buffer 22 and the frame buffer 23. A difference calculation circuit that calculates the difference of the difference images and generates a difference image; and 25, among the pixels of the difference image output from the difference calculation circuit 24, the pixel having a value exceeding the threshold set by the threshold calculation unit 31 is set to 1. , A threshold value processing circuit for generating a binary difference image in which other pixels are set to 0, and 26 is a binarized difference image generated by the threshold value processing circuit 25. Region shaping unit for shaping the frequency, 2
Reference numeral 7 is a labeling processing circuit for performing labeling of connected areas in the binarized difference image output from the area shaping unit 26 to generate a labeling image, and 28 is a labeling processing circuit 27.
Area position calculating means for calculating the position of each label area of the labeling image as the output of the label area in the monitoring area coordinates, and 29 is stored in the frame buffer 22 corresponding to each label area of the labeling image as the output of the labeling processing circuit 27. The feature amount such as the color and shape of the region of the frame image is calculated, and the feature amount and the position of each label region output from the region position calculating means 28 are processed and output as the moving region information of each moving region. The moving area information calculating unit 71 is a moving area information encoding circuit that encodes the moving area information output from the moving area information calculating unit 29 so that it can be transmitted by the LAN 6 in FIG.
Reference numeral 4 is a communication means that constitutes the LAN 6 for transmitting the moving area information coded by the moving area information coding circuit 71 to the moving object detecting means 3 of FIG. 1, and 72 is a moving area which is an output of the moving area information calculating section 29. Moving image processing evaluation means for outputting a moving area processing evaluation value which evaluates the calculation result of information, 30 is a moving area processing evaluation value output from the moving image processing evaluation means 72 and a moving object detection evaluation value decoding circuit 32. Moving image processing adjusting means for adjusting the state of moving region calculation based on the output of the moving object detection evaluation value, 3
Reference numeral 1 constitutes the moving image processing adjusting means 30. From the moving area processing evaluation value output from the moving image processing evaluation means 72 and the moving object detection evaluation value output from the moving object detection evaluation value decoding circuit 32. A threshold value calculation unit for calculating a threshold value used in the threshold value processing circuit 25, and 32 is a moving object that encodes the moving object detection evaluation value calculated by the moving object detection evaluation unit 5 in FIG. 1 and decodes the one transmitted by the LAN 6. Detection evaluation value decoding circuit, 35
Is a communication unit that constitutes the LAN 6 for transmitting the encoded moving object detection evaluation value. Further, 33 is the communication means described in FIG.

The operation of the moving image processing means 2 will be described below. The frame buffer 22 receives a moving image as a color digital video signal in the YUV format from the A / D converter 12 of FIG. 2 via the communication unit 21, and stores the image for one frame. The image for one frame stored in the frame buffer 22 is stored in the frame buffer 23 before the image for the next one frame is stored in the frame buffer 22.
Copy to.

The difference calculation circuit 24 uses the frame buffer 2
2 and the value of the Y signal corresponding to the luminance of each pixel of the continuous frame images stored in the frame buffer 23 is obtained to generate a difference image. The pixel that produces this difference is
Since it is the portion where the pixel values have changed between the two frame images, it basically corresponds to the moving object. The difference image calculated by the difference calculation circuit 24 is binarized by the threshold processing circuit 25. This binarization is performed by setting the pixel value of the pixel of the threshold value calculated by the threshold value calculator 31 to 1 when the pixel value of the difference image exceeds, and to 0 otherwise. Next, the binarized difference image output from the threshold processing circuit 25 is input to the area shaping unit 26 to shape the connected area. Hereinafter, the region shaping unit 26 will be described in more detail with reference to FIG.

FIG. 5 is a diagram showing the structure of the area shaping section 26 in this embodiment. In FIG. 5, 81 is an expansion processing circuit, 82 is a thinning processing circuit, and 83 is a noise removal circuit. The operation of the area shaping unit 26 will be described below. For the binarized difference image input from the threshold processing circuit 25, the dilation processing circuit 81 first performs dilation processing for setting pixels having a value of 0 among pixels adjacent to a pixel having a pixel value of 1 to 1. Apply once. Next, in the thinning processing circuit 82, thinning processing is performed three times to set pixels having a value of 1 among pixels adjacent to a pixel having a pixel value of 0 to 0. By the processing in the expansion processing circuit 81 and the thinning processing circuit 82, a small hole or a narrow crack in the connected region in which the value of the binary difference image is 1 is removed. In the binarized difference image that has been subjected to the expansion / thinning process, the noise removing circuit 83 then removes a small area. This process is a process of erasing a region in which the number of pixels is equal to or less than a predetermined value in the connected region included in the binarized difference image, that is, setting all pixel values to 0. By this processing, the connected region that does not correspond to the moving object due to noise or the like is removed. By the above processing,
The binarized difference image output from the threshold value processing circuit 25 of FIG. 3 is filled with holes and cracks in each connected region and the unnecessary connected region is removed by the region shaping unit 26.

The binarized difference image processed by the area shaping unit 26 is input to the labeling processing circuit 27 and subjected to labeling processing. In the labeling process in the present embodiment, a number is sequentially assigned to each connected region included in the input binarized difference image, and a pixel included in each connected region generates a labeling image in which the number of the connected region is set as a pixel value. By doing. In the following, this labeled connected area is referred to as a moving area.

The area position calculating means 28 calculates the position of the moving area corresponding to each label in the coordinates of the monitoring area from the labeling image output from the labeling processing circuit 27. This position can be regarded as the position of the moving body corresponding to each moving area. Hereinafter, the area position calculating means 28 will be described in more detail with reference to FIG.

FIG. 6 is a diagram showing the structure of the area position calculating means 28 in this embodiment. In FIG. 6, reference numeral 91 denotes a circumscribed rectangle calculation circuit that obtains a circumscribed rectangle of each connected area of the labeling image that is the output of the labeling processing circuit 27, and 92 corresponds to the circumscribed rectangle of each connected area that is the output of the circumscribed rectangle calculation circuit 91. A floor surface coordinate detection circuit for obtaining coordinates in an image of a portion where the moving object corresponding to the connected area contacts the floor surface, 93 is an output of the floor surface coordinate detection circuit 92, and the moving object corresponding to each connected area is the floor surface. It is a space coordinate calculating means for converting the coordinates in the contacting image into the coordinates of the monitoring area, and the circumscribing rectangle calculating circuit 91 and the floor surface coordinate detecting circuit 92 constitute the plane contact point coordinate calculating means 94.

Next, the operation of the area position calculating means 28 will be described. From the labeling image output from the labeling processing circuit 27, the circumscribing rectangle calculation circuit 91 obtains the circumscribing rectangle of each labeled connected region, and outputs the coordinates in the upper left and lower right images of the rectangle together with the label number. To do. In the floor surface coordinate detection circuit 92, the midpoint of the lower side of the circumscribed rectangle of each connected area, which is the output of the circumscribed rectangle calculation circuit 91, is defined as the coordinates in the image of the point at which the moving object corresponding to each connected area contacts the floor surface. Estimate and output the coordinate value together with the label of the corresponding motion area. In the spatial coordinate calculation means 93, when a plane is imaged by a camera, there is a relation of (Equation 1) and (Equation 2) between the point on the plane and the coordinate on the image corresponding to the point on the plane. Utilizing this, the coordinate value output from the floor coordinate detection circuit 92 is converted into a position on the floor surface, and the label corresponding to the coordinate value is output.

[0071]

[Equation 1]

[0072]

[Equation 2]

In (Equation 1) and (Equation 2), I
_x and I _y are the coordinates on the image of the point on the plane taken by the camera, X and Y are the coordinates of the coordinate system on the plane of the point on the plane,
a ₁ , a ₂ , b ₁ , b ₂ , b ₃ , c ₁ , c ₂ , c ₃ are parameters that depend on the relative relationship between the camera and the plane, the floor coordinate system, and the way the camera image coordinate system is taken. , If the camera is fixed, it can be calculated in advance by the regression method by measuring the coordinates on the image of some known points. As described above, the area position calculation means 28 calculates the position in the monitored area coordinates of the moving object corresponding to the labeled connected area in the labeling image which is the output of the labeling processing circuit 27, together with the corresponding label. Output.

The moving area information calculation unit 29 outputs a frame image output from the frame buffer 22, a labeling image output from the labeling processing circuit 27, and an animal corresponding to each moving area output from the area position calculating means 28. The position of the body is input, and the moving area corresponding to the moving object in the monitoring area and the moving area information including the feature amount thereof are calculated and output. In this embodiment, the moving area information is the label number of each moving area, the image coordinates of the upper left and lower right of the circumscribed rectangle of each moving area, the representative chromaticity of the moving area, and the moving object corresponding to each moving area. And the height of the moving body corresponding to each moving area. Hereinafter, the moving area information calculation unit 29 will be described in more detail with reference to FIG. 7.

FIG. 7 is a diagram showing the structure of the moving area information calculation section 29 together with related parts. In FIG. 7, 10
Reference numeral 1 denotes a circumscribing rectangle calculation circuit that obtains a circumscribing rectangle of each connected area of the labeling image output from the labeling processing circuit 27 and outputs image coordinates of the upper left and lower right corners of the rectangle. 102 indicates a frame image output from the frame buffer 22. Area color calculation means for calculating the representative chromaticity of each moving area from the labeling image output from the labeling processing circuit 27, and 103 is the image coordinates output from the circumscribing rectangle calculation circuit 101 and the output from the area position calculation means 28. Area shape calculation means for calculating the height of the moving body corresponding to each moving area from the position of the moving body corresponding to each moving area, 104 is the coordinates of the vertices of the circumscribing rectangle which is the output of the circumscribing rectangle calculating circuit 101, Area color calculation means 102
Is a moving area information generating unit that collects the representative chromaticity that is the output of the area shape calculating unit 103, the object height that is the output of the area shape calculating unit 103, and the position in the monitoring area coordinates as the moving area information. A moving area information calculating unit 29 is configured by the color calculating unit 102, the area shape calculating unit 103, and the moving area information generating unit 104.

Next, the operation of the moving area information calculation unit 29 will be described. The circumscribing rectangle calculation circuit 101 obtains a circumscribing rectangle whose sides are parallel or perpendicular to each axis of the image coordinates of each moving area of the labeling image output from the labeling processing circuit 27, and coordinates of the upper left vertex and the lower right vertex. Is output together with the label of each moving area. The area color calculation unit 102 generates a histogram of each of the U and V pixel values in the frame image of the frame buffer 22 of the pixel corresponding to each moving area of the labeling image output from the label processing circuit 27, and becomes a peak. The values of U and V are output as the representative chromaticity. The area shape calculation means 103 calculates the height of the circumscribed rectangle from the vertex coordinates of the circumscribed rectangle output from the circumscribed rectangle calculation circuit 101, and the height of the circumscribed rectangle corresponds to the output of the area position calculation means 28. The height of the moving object corresponding to each moving area is calculated from the position information of the moving object and the focal length and position of each camera stored in advance, and the position information and the height of the moving object are output together with the label. . Motion area information generation unit 10
Reference numeral 4 indicates the coordinates of the vertices of the circumscribed rectangle that is the output of the circumscribed rectangle calculation circuit 101, the representative chromaticity that is the output of the region color calculation unit 102, the object height that is the output of the region shape calculation unit 103, and the monitoring area coordinates. The moving area information composed of the position in and the corresponding label is generated and output for each moving area. As described above, the moving area information calculation unit 29 receives the outputs of the labeling processing circuit 27, the area position calculating means 28, and the frame buffer 22, and outputs the moving area information.

In FIG. 3, a moving area information coding circuit 71 is provided.
1 encodes the moving area information so that it can be transmitted by the LAN 6 of FIG. 1, and sends the moving area information to the moving object detecting means 3 via the communication means 34.

The moving image processing evaluation means 72 receives the moving area information output from the moving area information calculating section 29, and if the height of the moving body corresponding to each moving area is smaller than the preset range, 1 is output to the threshold value calculation unit 31 of the moving image processing adjustment unit 30 as a moving area processing evaluation value when 1 is large, 1 when it is within the range, and 0 when it is within the range. The height range set in advance here is the range of the height of the moving object to be monitored by the distributed monitoring apparatus according to the present embodiment, and the height of the moving object corresponding to each moving region is The fact that it is smaller than the range corresponds to the fact that the moving region is detected smaller than it actually is because the threshold for binarization by the threshold processing circuit 25 is too large. On the contrary, the height of the moving object corresponding to each moving area is larger than the above range, which means that the threshold value for binarization is too small and the moving area is detected larger than it actually is. To do.

The moving object detection evaluation value decoding circuit 32 receives the encoded evaluation value of the moving object information from the moving object detection means 3 of FIG. 1 via the communication means 35, decodes it, and decodes it. The detection evaluation value is output to the threshold value calculation unit 31 of the moving image processing adjustment unit 30 and is also sent to the aperture adjustment amount calculation circuit 15 of FIG. 2 via the communication unit 33. As described above, the animal body detection evaluation value in the present Example is a value indicating the degree of whether the animal body is excessively detected as a result of the detection of the animal body, is under-detected, or is detected in an appropriate number. is there.

Threshold value calculating section 31 of moving image processing adjusting means 30
Receives as input the moving area processing evaluation value output from the moving image processing evaluation means 72 and the moving object detection evaluation value output from the moving object detection evaluation value decoding circuit 32, and sets the threshold value used in the threshold value processing circuit 25. calculate. The method of calculating this threshold value follows the following algorithm.

1) If no past threshold is set, such as at the start of processing, the default value is used as a temporary threshold, and in other cases, the past threshold is used as a temporary threshold.

2) When the evaluation value of the moving body detection indicates that the moving body is under-detected, the provisional threshold value is decreased by a preset value to detect the moving body excessively. If it indicates that the provisional threshold value is increased by a preset value, and if it indicates that an appropriate number of detections have been made, the provisional threshold value is left unchanged.

3) Referring to the value of the moving area processing evaluation value, when the moving area processing evaluation value is 1, the temporary threshold value is increased by a preset value, and when it is -1, the temporary threshold value is set. Is decreased by a preset value, and when it is 0, the temporary threshold value is left as it is.

As described above, the moving image processing adjusting means 3
The threshold value calculation unit 31 of 0 changes the threshold value of the threshold value processing circuit 25 according to the result of the moving area calculation in the moving image processing unit 2 and the evaluation of the result of the moving object detection by the moving object detection unit 3, Change the calculation state of the moving area.

As described above, the moving image processing means 2 in this embodiment changes the moving region calculation state according to the evaluation of the moving region calculation result and the moving object detection result, and the moving image of FIG. The moving area information at that time can be calculated and output from the color digital video signal of the YUV format input from the input means 1.

(Moving Object Detecting Means) Next, the moving object detecting means 3 of FIG. 1 will be described with reference to FIG. 4, 34 is included in the LAN 6 of FIG. 1 for receiving the encoded moving area information output from the moving area information encoding circuit 71 from the moving image processing means 2 and 8, as described in FIG. Communication means 52 is a moving area information decoding circuit which decodes the coded moving area information sent via the communication means 34 and outputs it as moving area information. 53 is an output of the moving area information decoding circuit 52. A feature association unit 54 that associates each moving region with a moving object detected in the past from certain moving region information and estimated moving object information output from the moving object information estimating unit 57.
Is an unsupported moving object processing unit that estimates the reason why a moving object detected in the past that has not been matched by the feature matching unit 53 has not been matched, and performs processing in accordance therewith.
Reference numeral 5 is an uncorresponding moving region processing unit that estimates the reason why the moving region that has not been correlated with the moving object detected in the past by the feature correlation unit 53 is not correlated, and performs processing in accordance therewith.
Reference numeral 56 is an association determination unit that determines the association between the moving region and the previously detected moving object from the output of the unsupported moving object processing unit 54 and the output of the unsupported moving object processing unit 55, and 57 is the moving object information storage unit. The moving body information estimating means for estimating the moving body information at the present time from the moving body information stored in 4 in FIG. 4, 58 is the result of associating the moving region output from the association determining unit 56 with the previously detected moving body. From the above, an object body information calculation unit that determines the state of the object body in the monitored area and outputs it as the object body information, 61 is a past object body information stored in the object body storage unit 4 and a body object information calculation unit 58. A moving object detection evaluation circuit for calculating an evaluation value of the detection state of the moving object of the moving object detecting means 3 from the current moving object information which is an output, and 62 is a moving object detection output of the moving object detection evaluation circuit 61. Evaluation value can be transmitted by LAN6 As described with reference to FIG. 3, the moving object detection evaluation value encoding circuit 35, which encodes as described above, constitutes the LAN 6 for transmitting the moving object detection evaluation value encoded by the moving object detection evaluation value encoding circuit 62. It is a communication means. In addition,
The feature association unit 53, the unsupported moving object processing unit 54, the unsupported moving region processing unit 55, and the association determination unit 56 constitute the moving object moving region association unit 51. Further, the moving body information decoding circuit 52, the moving body moving area associating means 51, the moving body information calculating section 58, and the moving body information estimating means 57 constitute the moving body detecting means 3. Further, the moving object detection evaluation circuit 61 and the moving object detection evaluation value encoding circuit 62 constitute the moving object detection evaluation means 5.

Next, the operation of the moving object detecting means 3 will be described. The moving area information decoding circuit 52 outputs the moving area information, which is the output of the moving area information encoding circuit 71 of FIG.
It is received from the moving image processing means 2 and 8, and is decoded and output as moving area information. The feature associating unit 53 of the moving object associating unit 51 of the moving object has the moving region information decoding circuit 5
2 from the moving area information and the moving object information estimating means 57 output from the previously detected estimated moving object information of the moving object, each moving area calculated by the moving image processing means 2 and 8 and the past moving object information. The detected moving bodies are associated with each other. Hereinafter, each part of the animal body motion area associating means 51 will be described in more detail.

((Animal body moving area associating means)) ((feature associating portion)) FIG. 8 is a diagram showing the configuration of the feature associating portion 53 in the animal body moving area associating means 51. In FIG. 8, reference numeral 111 refers to moving area information and estimated moving object information of a moving object detected in the past, and associates a moving area with a moving object detected in the past by distance, and the moving area information that is an input of the association candidate. A shape correspondence evaluation indicating the degree of correspondence between the moving regions matched by the distance correspondence circuit 111 and the shape of the matching candidate of the moving body detected in the past. A shape matching circuit that calculates a value, 113 is a chromaticity matching circuit that calculates a chromaticity correspondence evaluation value that indicates the degree of correspondence by the representative chromaticity of the correspondence candidate that is the output of the shape correspondence circuit 112, and 114 is a color It is a correspondence determination circuit that determines a final correspondence from the shape correspondence evaluation value, the chromaticity correspondence evaluation value, and the correspondence candidate output from the degree correspondence circuit 113.

In the present embodiment, the moving area information corresponds to the label number of each moving area, the image coordinates of the upper left and lower right of the circumscribing rectangle, and the representative chromaticity of the moving area, as described above. It is composed of the position of the moving body and the height of the corresponding moving body. The label numbers of the moving area information input from the moving image processing means 2 and the moving image processing means 8 do not overlap.

The moving body information includes the number of the moving body, the representative chromaticity of the moving body, the position of the moving body in the coordinates of the monitoring area,
From the position of the moving object in the image coordinate system of the moving image captured by the moving image input means 1, the position of the moving object in the image coordinate system of the moving image captured by the moving image input means 7, and the height of the moving body Composed. Further, the moving area information and the moving object information are associated with the moving object information of the moving image processing means 2 and 8 in FIG. 1 separately from the estimated moving object information. That is, the correspondence between the moving area and the moving object detected in the past is independently performed by the outputs of the moving image processing means 2 and 8 in FIG. Hereinafter, a collection of moving area information of moving areas calculated by the moving image processing means 2 will be referred to as moving area information A, and a collection of moving area information of moving areas calculated by the moving image processing means 8 will be referred to as moving area information B. .

The operation of the feature association unit 53 will be described below. To the distance association circuit 111, the moving area information A and B are input from the moving area information decoding circuit 52 of FIG. 4, and the estimated moving object information of the moving object detected in the past is input from the moving object information estimating means 57. Correspondence by distance is performed in the following steps.

1) Of the moving objects detected in the past, the estimated moving object information corresponding to the moving object that has not yet been associated is selected.

2) The following steps 3 to 5 are performed on each of the moving area information A and the moving area information B.

3) The Euclidean distance between the estimated position of the currently selected moving object in the monitoring area and the position of the moving object in the monitoring area corresponding to the moving area in each moving area information is calculated.

4) Of the moving areas, those having the Euclidean distance calculated in step 3 larger than a preset threshold value are removed from the candidates for association with the currently selected moving object.

5) The label numbers of the remaining moving areas are arranged in ascending order of Euclidean distance calculated in step 3,
This is a candidate for association with the currently selected moving object.

6) If there is a moving object that has not been associated, the process returns to step 1; otherwise, the process ends.

By the processing by the above steps, the moving area information A for each moving object previously detected by the distance associating circuit 111.
And the moving area are associated with the moving area information B, respectively.
However, one moving object does not always correspond to one moving area based on the moving area information A and the moving area information B, and there may be a case where there is no corresponding moving area or a plurality of corresponding moving areas. Further, it may be associated with the moving area of the moving area information A and may not be associated with the moving area of the moving area information B. Then, the distance association circuit 111 outputs the input moving region information A and B and the estimated moving body information together with the moving region association candidate for each moving object.

The distance associating circuit 111 of the feature associating unit 53 in this embodiment uses the Euclidean distance, but the present invention is not limited to this, and for example, a weighted Euclidean distance may be used.

Next, the shape association circuit 112 receives the output of the distance association circuit 111, which is the candidate for association of the moving area with respect to the moving object detected in the past, the moving area information A and B, and the estimated moving object information. The processing is performed in the following steps.

1) The following steps 2 and 3 are performed for each moving object detected in the past.

2) The difference between the height of the moving area of each association candidate and the estimated height of the moving object is calculated.

3) The shape correspondence evaluation value is set to 0 for the moving area of the association candidate whose difference calculated in step 2 exceeds the preset threshold value, and the moving area of the association candidate below the preset threshold value is the shape. The corresponding evaluation value is 1.

As described above, the shape matching circuit 11
2, the height of the moving body detected in the past and the moving body corresponding to the moving region is used as shape information, and the difference is used as an index of similarity,
Calculate the shape correspondence evaluation value. Then, the association candidate of the moving area with respect to the moving object detected in the past to which the shape-corresponding evaluation value is added, the moving area information A and B, and the estimated moving object information are output.

The shape correspondence circuit 112 of the present embodiment.
Uses the height of the animal body as the shape, but is not limited thereto, the width of the animal body, the aspect ratio of the animal body,
Other shape features such as the circularity of the moving body may be used.

Next, the chromaticity associating circuit 113 associates the moving object information and the moving object information A, B with the previously detected moving object and moving object associating candidates to which the shape corresponding evaluation value output from the distance associating circuit 111 is added. The estimated moving object information is input and the processing is performed in the following steps.

1) The following steps 2 and 3 are performed for each moving object detected in the past.

2) The Euclidean distance between the representative chromaticity of the moving area of each association candidate and the estimated representative chromaticity of the moving object is calculated.

3) The moving region of the association candidate in which the Euclidean distance calculated in step 2 exceeds the preset threshold has the chromaticity correspondence evaluation value set to 0, and the movement of the association candidate below the preset threshold is set. The region has a chromaticity correspondence evaluation value of 1.

In this way, the chromaticity matching circuit 113
Is the representative chromaticity of the moving object corresponding to the moving object and the moving object detected in the past, and the Euclidean distance between them is used as an index of the color similarity between the moving object and the moving object. The chromaticity correspondence evaluation value is calculated for each correspondence candidate of. Then, it outputs the previously detected association candidates of the moving body and the moving area, the moving area information A and B, and the estimated moving body information to which the shape-based evaluation value and the chromaticity-based evaluation value are added.

The chromaticity matching circuit 113 of this embodiment is used.
Uses the representative chromaticity of the peak of the histogram of U and V values as a color, but is not limited to that. For example, a and b of the Lab color system and the saturation and hue obtained from them are used as colors. You may use.

Next, the association determining circuit 114 associates the previously detected moving object with the moving object detected by adding the shape corresponding evaluation value and the chromaticity corresponding evaluation value, the moving area information A and B, and the estimated moving object information. Using as input, at most one moving area corresponding to each moving object detected in the past is determined from the moving area information A and moving area information B in the following steps.

1) The following steps 2 to 4 are performed for each moving object detected in the past. 2) The following steps 3 and 4 are performed for each moving area candidate of moving area information A and moving area candidate of moving area information B. 3) The shape-corresponding evaluation value and the chromaticity-corresponding evaluation value are added for each moving region candidate to obtain a correspondence evaluation value. 4) The corresponding evaluation values of the moving area candidates are compared, and if there is one moving area having the maximum value, it is determined that the moving area corresponds to the moving object. When there are a plurality of moving areas having the maximum value, it is determined that the first moving area in the order arranged by the distance association circuit 111 corresponds to the moving object.

As described above, the feature association unit 53
Is the distance, shape in the coordinate system in the real space of the monitored area,
In consideration of chromaticity and the like, at most one moving area corresponding to the moving object detected in the past is determined from the moving area information A and the moving area information B. Then, together with this association result, the input motion area information A and B and the estimated moving object information of the moving object detected in the past are output.

((Unsupported moving object processing unit)) The unsupported moving object processing unit 54 shown in FIG. 4 is the moving region information as a result of the process of associating the moving region in the feature associating unit 53 with the moving object detected in the past. A moving object in which one moving area is not associated with both A and moving area information B is processed. Hereinafter, the unsupported moving object processing unit 54 will be described in more detail with reference to FIG. 9.

FIG. 9 is a diagram showing the structure of the unsupported moving object processing unit 54. In FIG. 9, 121 is a moving image input means 1
And a concealment object position storage means for storing the position of the stationary concealment object in each surveillance area of the moving image input means 7 and 12
Reference numeral 3 is a moving image input means position storage means for storing the positions of the moving image input means 1 and the moving image input means 7 in the surveillance area coordinates, and 122 is the position of the stationary concealment object stored in the concealment object position storage means 121. Based on the position of the moving object detected in the past and the positions of the moving image inputting means 1 and the moving image inputting means 7 stored in the moving image inputting means position storing means 123, the moving body detected in the past is the moving image inputting means 1. Alternatively, a moving object inputting means 7 determines whether or not the moving object is hidden by a stationary concealing object, and 124 is a moving image input by a moving object detected in the past by another moving object detected in the past. The moving object input means 1 or moving image input means 7 determines whether or not the moving object is hidden by the moving image input means 7.
Alternatively, a photographing range determination unit that determines whether the moving image input unit 7 is within the photographing range, and 126 is a distance threshold value storage that stores a distance threshold value that indicates a moving object detection range of the moving image input unit 1 and the moving image input unit 7. Means 127 is a moving object detected in the past from the positions of the moving image input means 1 and the moving image input means 7 stored in the moving image input means position storage means 123 and the distance threshold value stored in the distance threshold value storage means 126. It is a moving area detection range judging means for judging whether or not the moving area is close enough to the moving image input means 1 or the moving image input means 7.

The operation of the unsupported moving object processing unit 54 is as follows. The concealed object position storage means 121 stores in advance the positions of the moving picture input means 1 and the moving picture input means 7 in the surveillance area coordinates of the stationary concealed object within the field of view of the television camera. Further, the positions of the moving image input means 1 and the moving image input means 7 in the coordinates of the monitoring area are stored in advance in the moving image input means position storage means 123.

The concealment judging means 122 includes the feature associating unit 5
The result of associating the moving area with the moving object detected in the past, the moving area information A, the moving area information B, and the estimated moving object information of the moving object detected in the past are input from 3 and are processed in the following steps. .

1) Of the moving objects detected in the past, those having no corresponding moving area are selected in order from the moving area information A, and the process of step 2 is performed.

2) A line segment connecting the position of the moving image input unit 1 stored in the moving image input unit position storage unit 123 and the position of the selected moving object is calculated, and the straight line and the concealed object position storage unit are calculated. If the distance between each concealed object position stored in 121 is within a preset threshold value, it is determined that the object is concealed by the concealed object, and the moving image processing means 2 determines The empty moving area is made to correspond to the moving area corresponding to.

3) Of the moving objects detected in the past, those having no corresponding moving area are selected in order from the moving area information B, and the process of step 4 is performed.

4) Moving image input means: A line segment connecting the position of the moving image input means 7 stored in the position storing means 123 and the position of the selected moving body is calculated, and the straight line and the concealed object position storing means are calculated. If the distance between each concealed object position stored in 121 is within a preset threshold value, it is determined that the object is concealed by the concealed object, and the moving image processing means 8 determines The empty moving area is made to correspond to the moving area corresponding to. In the above step,
The moving area information of an empty moving area is information in which each possible feature value is set to an impossible value that is determined in advance.

In this way, the concealment judging means 122 in this embodiment makes a proper reasoning when the moving object detected in the past is concealed by the stationary concealing object and associates it with the empty moving area. Then, the association result of the moving area and the moving object detected in the past, the moving area information A, the moving area information B, and the estimated moving object information of the moving object detected in the past are output.

The concealment judging means 122 in the moving body motion area associating means 51 of the present embodiment uses the distance of the stationary concealing object from the line segment connecting the estimated position of the moving body detected in the past and the television camera, Although the concealment is determined, the concealment range is not limited to that. For example, it is presumed that the concealment range of the stationary concealment object is accurately obtained in advance from the position of the TV camera, and it is estimated that the moving object is within the range. In some cases, it may be determined that it is hidden.

Next, the moving object concealment judging means 124 associates the moving area with the moving object detected in the past from the concealment judging means 122, moving area information A and B, and the estimated animal of the moving object detected in the past. Body information is input and processing is performed in the following steps.

1) Of the moving objects detected in the past, those having no corresponding moving area are selected in order from the moving area information A, and the process of step 2 is performed.

2) A line segment connecting the position of the moving image input unit 1 stored in the moving image input unit position storing unit 123 and the position of the selected moving object is calculated, and the corresponding line segment is selected from the moving region information A. The distance to the estimated position of the moving object detected in the past where there was a moving area is calculated. If there is a moving object whose distance is less than or equal to a preset threshold, it is determined that the selected moving object is hidden by the moving object, and an empty moving area is set as the corresponding moving area. To do.

3) Of the moving objects detected in the past, those having no corresponding moving area are selected in order from the moving area information B, and the process of step 4 is performed.

4) A line segment connecting the position of the moving image input unit 7 stored in the moving image input unit position storage unit 123 and the position of the selected moving object is calculated, and the corresponding line segment is selected from the moving region information B. The distance to the estimated position of the moving object detected in the past where there was a moving area is calculated. If there is a moving object whose distance is less than or equal to a preset threshold, it is determined that the selected moving object is hidden by the moving object, and an empty moving area is set as the corresponding moving area. To do.

As described above, the moving object hiding judging means 1
24 correctly determines the reason why the object cannot be associated when the other object hides another object, and in that case, the empty moving area is associated. Then, the association result of the moving area and the moving object detected in the past, the moving area information A and B, and the estimated moving object information of the moving object detected in the past are output.

Although the moving object hiding judging means 124 in this embodiment judges the hiding by the moving object by using the distance between the line segment connecting the television camera and the moving object and the other moving object, it is not limited to this. However, for example, whether the image coordinates in the moving image of the corresponding moving image input means of the uncorresponding moving object is in the moving region corresponding to the already matched moving object. Concealment by the moving body may be determined.

Next, the photographing range judging means 125 estimates the moving area and the moving object detected in the past by the moving object hiding judging means 124, moving area information A and B, and estimates the moving object detected in the past. The moving body information is input and processing is performed in the following steps.

1) Of the moving objects detected in the past, those having no corresponding moving area are selected in order from the moving area information A, and the process of step 2 is performed.

2) When the coordinates of the moving object in the image coordinate system of the moving image picked up by the moving image input means 1 of the selected moving object information is outside the image range of the moving image of the moving image input means 1. Then, it is determined that the selected moving object is outside the imaging range of the moving image input means 1, and the empty moving region is set as the corresponding moving region.

3) Of the moving objects detected in the past, those having no corresponding moving area are sequentially selected from the moving area information B, and the process of step 4 is performed.

4) When the coordinates of the moving object in the image coordinate system of the moving image picked up by the moving image input means 7 of the selected moving object information is outside the image range of the moving image of the moving image input means 7. Then, it is determined that the selected moving object is outside the imaging range of the moving image input means 7, and the empty moving region is set as the corresponding moving region.

As described above, the photographing range determining means 12
5 uses the coordinates of the moving object in the image coordinate system of the moving image input means to correctly determine the reason why the moving object cannot be matched when it is outside the image capturing range of the television camera, and matches the empty moving region. Then, the association result of the moving area and the moving object detected in the past, the moving area information A and B, and the estimated moving object information of the moving object detected in the past are output.

In the distance threshold storage means 126, a distance that allows an assumed moving object to be imaged with a size that is surely detected as a moving area is calculated in advance and stored as a distance threshold in advance, corresponding to each moving image input means. I'll do it.

Next, the moving area detection range judging means 127 associates the moving area with the moving object detected in the past from the photographing range judging means 125, moving area information A and B, and the moving object detected in the past. Estimated moving object information is input and processing is performed in the following steps.

1) Of the moving objects detected in the past, those having no corresponding moving area are selected in order from the moving area information A, and the processes of steps 2 and 3 are performed.

2) Moving image input means 1 stored in the selected moving object position and moving image input means position storing means 123
Find the Euclidean distance from the position.

3) When the distance obtained in step 2 exceeds the distance threshold corresponding to the moving image input means 1 stored in the distance threshold storage means 126, the selected moving object is the moving image input means 1 It is determined that the user is outside the moving object detection range, and an empty moving area is set as the corresponding moving area.

4) Of the moving objects detected in the past, those having no corresponding moving area are selected in order from the moving area information B, and the processes of steps 5 and 6 are performed.

5) Moving image input means 7 stored in the selected moving object position and moving image input means position storing means 123
Find the Euclidean distance from the position.

6) When the distance obtained in step 5 exceeds the distance threshold value corresponding to the moving image input means 7 stored in the distance threshold storage means 126, the selected moving object is the moving image input means 7 It is determined that the user is outside the moving object detection range, and an empty moving area is set as the corresponding moving area.

In this way, the moving area detection range determination means 12
7 uses the distance between the TV camera and the moving object in the surveillance area coordinates to correctly explain the reason why the moving object cannot be associated when it is too far from the TV camera to capture a moving area. Judgment is made and the sky moving area is associated.

As described above, the unsupported moving object processing unit 5
4 is outside the imaging range of the moving image input means 1 or the moving image input means 7 when the moving object detected in the past is hidden by a stationary object, when it is hidden by another moving object. In the case where the moving object is outside the moving object detection range of the moving image input means, the reason why the previously detected moving object cannot be associated with the moving area is determined. Is added to the association information that is the output of the feature association unit 53, and is output together with the moving area information A, the moving area information B, and the estimated moving object information of the moving object detected in the past. To do.

((Unsupported Moving Area Processing Unit)) The unsupported moving area processing unit 55 of FIG. 4 performs the processing of the moving area which is not associated with the moving object detected in the past by the feature association unit 53 as follows. Follow the steps. Hereinafter, among the moving areas of the moving area information A, the moving areas that are not associated with the moving object detected in the past are referred to as uncorresponding moving areas A, and the moving areas of the moving area information B that have been detected in the past. A moving area that is not associated with a moving object is called an unsupported moving area B.

1) Each moving area of the non-corresponding moving area A is sequentially selected, and the process of step 2 is performed.

2) The distance between the position of the moving body corresponding to the selected moving area and the position of the moving body corresponding to each moving area of the non-corresponding moving area B is calculated, and the distance is set to a preset threshold value. If there is the following and the smallest one, the label number of the moving area of the unsupported moving area B is temporarily associated with the selected moving area of the unsupported moving area A.

3) Each moving area in the uncorresponding moving area B is sequentially selected, and the process of step 4 is performed.

4) The distance between the position of the moving body corresponding to the selected moving area and the position of the moving body corresponding to each moving area of the non-corresponding moving area A is calculated, and the distance is set to a preset threshold value. If there is the following and the smallest one, the label number of the moving area of the uncorresponding moving area A is provisionally associated with the selected moving area of the uncorresponding moving area B.

5) The label number of the moving area temporarily associated with the unsupported moving area B temporarily associated with each moving area of the unsupported moving area A is the moving area of the original unsupported moving area A. In the case of the label number, that is, when the motion area of the non-corresponding motion area A and the motion area of the non-corresponding motion area B are provisionally associated with each other, new empty moving object information is generated, and The moving areas of the corresponding moving area A and the non-corresponding moving area B are associated as corresponding moving areas.

6) New moving object information is generated for each of the non-corresponding motion area A and the non-corresponding motion area B which are not associated with the moving object in step 5, and each motion area is defined as follows. It is associated as a moving area corresponding to the moving body. In this case, only one moving region corresponds to one moving object, but it is left as it is.

As described above, the unsupported moving area processing unit 5
Reference numeral 5 is for newly generating moving object information from moving areas that have not been associated with moving objects detected in the past by the feature associating unit 53, and associating the moving areas with the moving object information, and associating these moving object information. Output with.

((Association Confirmation Unit)) The association determination unit 56 of FIG.
The moving area information B, the estimated moving object information of the moving object detected in the past, and the matching result of the moving object detected in the past are input. The moving object information newly generated from the moving area that is not associated with the previously detected moving object and the association information thereof are input. Then, in the processing by the following steps, the association between the moving object detected in the past and the moving area is determined and output as moving object information.

1) The correspondence information of the moving body and the moving area inputted from the unsupported moving body processing section 54 and the correspondence information of the moving body and the moving area inputted from the unsupported moving body processing section 55 are combined. It collects in one correspondence information.

2) Of the moving object information, the moving object information A or moving object information B, which has no corresponding moving object information, is removed.

Through the above processing, the association confirmation unit 56
Determines the correspondence between the moving object and the moving area detected in the past, and newly rearranges the matching information, moving area information A, moving area information B, and estimated moving object information of the moving object detected in the past.
The generated moving object information is output.

As described above, the moving body moving area associating means 51 detects the moving body detected in the past and the moving image processing means 2.
And the moving area information calculated by the moving image processing means 8 are associated with each other, and the moving object that is not associated is forcibly dealt with by associating the corresponding empty moving area with the reason. A moving object that is not associated with the moving object is newly generated and associated, and moving object information and moving area information A and moving area information are associated with the moving object. B, the estimated moving object information of the moving object detected in the past, and the newly generated moving object information are output.

((Animal body information calculating unit)) The moving body information calculating unit 58 in FIG. 4 outputs the moving body information and the moving body correspondence information output from the moving body moving area associating means 51.
The moving object information B, the estimated moving object information of the moving object detected in the past, and the newly generated moving object information are input, and the past moving object information stored in the moving object information storage unit 4 is referred to. While calculating the current moving object information. Then, the moving body information is output to the outside as an output of the distributed monitoring apparatus according to the present embodiment, and is output to the moving body information storage unit 4 and the moving body detection / evaluation circuit 61. It should be noted that the position of the moving object in the image coordinate system of the moving image input means 1 and the moving image input means 7 in the moving body information is not necessary as an output of the moving body information calculation unit 58 in the present embodiment, and thus new calculation is not performed. Not performed. Hereinafter, the moving object information calculation unit 58 will be described in more detail with reference to FIG.

FIG. 10 is a diagram showing the structure of the moving object information calculation unit 58. In FIG. 10, 131 is a moving body number determining unit that determines the number of each moving body, 132 is a moving body position determining unit that calculates the position of each moving body in the coordinates of the monitored area, 133
Is a moving object shape determining unit that calculates the height of each moving object, 135
Is a body color determination unit that calculates the representative chromaticity of each body, 13
Reference numeral 6 denotes a moving object information determination unit that outputs moving object information collectively as an output of the distributed monitoring apparatus according to this embodiment.

Next, the operation of the moving object information calculation section 58 will be described. The moving object number determining unit 131 estimates moving object information, moving area information A and moving area information B, which is output from the moving object moving area associating unit 51, and moving object information A and moving area information B. Animal body information and newly generated animal body information are input. Then, while referring to the past moving body information stored in the moving body information storage unit 4, the number of the estimated moving body information of the previously detected moving body output from the unsupported moving body processing unit 54 is not changed. A number other than the number used in the past moving object information stored in the moving object information storage means 4 is assigned to the newly generated moving object information from the smallest. In other words, a number that is not used in a previously detected moving object is assigned to the newly generated moving object information in ascending order. Then, the estimated moving body information of the moving body detected in the past and the newly generated moving body information are combined into one as provisional moving body information, and the correspondence information between the moving body and the moving area, and the moving area information. It is output together with A and moving area information B. Note that the numbers do not have to be consecutive after the numbers are assigned, for example, because the number of the numbers removed by the association confirming unit 56 is larger than that of the newly generated moving object information.

Next, the moving body position deciding unit 132 receives the information on the correspondence between the moving body and the moving area from the moving body number deciding unit 131,
The moving area information A, the moving area information B, and the numbered provisional moving object information are received, and the position of the moving object in the monitoring area corresponding to each moving object information according to the algorithm classified in each case as follows. To calculate. In each of the following equations (3) to (5), P _new is a new position in the surveillance zone coordinates of the moving object, P _old is an estimated position in the surveillance zone coordinates of the previously detected _moving object, and α is It is a constant of 0 or more and 1 or less.

1) Moving area information A for moving objects detected in the past
If the corresponding moving area exists from both of the moving area information B and the moving area information B, the new position in the monitored area of the moving object detected in the past is determined according to (Equation 3).

[0166]

[Equation 3]

In the equation (3), p _A is the position of the corresponding moving area of the moving area information A in the monitoring area coordinates, and p _B is the position of the corresponding moving area of the moving area information B in the monitoring area coordinates.

2) Moving area information A for moving objects detected in the past
Alternatively, if only one of the moving area information B is associated with the moving area and the other is associated with the empty moving area, the monitored area of the moving object detected in the past according to (Equation 4). Determine the new position in coordinates.

[0169]

[Equation 4]

In (Equation 4), p is the position of the non-empty moving region corresponding to the moving object.

3) When the moving object information is new moving object information generated by the unsupported moving area processing unit 55 and there is a moving area corresponding to both moving area information A and moving area information B, ( The position of the newly generated moving body in the monitored area coordinates is determined according to the equation (5).

[0172]

[Equation 5]

In (Equation 5), p _A is the position of the moving area corresponding to the moving object in the moving area information A, and p _B is the position of the moving area corresponding to the moving object in the moving area information B.

4) If the moving object information is newly generated by the unsupported moving area processing unit 55 and the moving area corresponds only to either one of the moving area information A and the moving area information B, The position of the corresponding moving area is set as the position of the newly generated moving object.

In accordance with the above algorithm, the moving body position determining unit 132 calculates the position of each provisional moving body information in the monitored area coordinates of the moving body, and associates the moving body with the moving area, the moving area information A, and the moving area information A. The area information B and the provisional moving object information in which the position of each moving object in the monitored area coordinates is updated are output.

Next, the moving object shape determining unit 133 sends the moving object position determining unit 132 information on the correspondence between the moving object and the moving area,
The moving area information A, the moving area information B, and the provisional moving object information are received, and the height of the moving object corresponding to each moving object information is calculated according to the algorithm classified according to the case as follows. In each of the following equations (6) to (8), H _new is the new height of the moving body, H _old is the estimated height of the moving body detected in the past, and α is a constant of 0 or more and 1 or less. Is.

1) Moving area information A for moving objects detected in the past
If the corresponding moving area exists from both the moving area information B and the moving area information B, the new height of the moving object detected in the past is determined according to (Equation 6).

[0178]

[Equation 6]

In (Equation 6), h _A corresponds to the height of the moving object corresponding to the moving object in the moving area information A, and h _B corresponds to the moving area corresponding to the moving object in the moving area information B. The height of the moving body.

2) Moving area information A for moving objects detected in the past
Alternatively, when only one of the moving area information B is associated with the moving area and the other is associated with the empty moving area, the monitoring area of the moving object detected in the past according to (Equation 7) is used. Determine a new position in.

[0181]

[Equation 7]

In (Equation 7), h is the height of the moving body corresponding to the non-empty moving region corresponding to the moving body.

3) The moving object information A and the moving area information B are added to the moving object information newly generated by the unsupported moving area processing unit 55.
If there is a corresponding motion area from both, (Equation 8)
The position of the newly created moving body in the monitored area is determined in accordance with.

[0184]

[Equation 8]

In (Equation 8), h _A corresponds to the height of the moving object corresponding to the moving object in the moving area information A, and h _B corresponds to the moving area corresponding to the moving object in the moving area information B. The height of the moving body.

4) The moving area information A or the moving area information B is added to the moving object information newly generated by the unsupported moving area processing unit 55.
When there is a corresponding moving area from only one of the above, the height of the moving object corresponding to the corresponding moving area is set as the height of the newly generated moving object.

According to the above algorithm, the moving object shape determining unit 133 calculates the height of the moving object corresponding to each provisional moving object information, and associates the moving object with the moving area, the moving area information A, and the moving area. The information B and the provisional moving object information in which the height of each moving object is updated are output.

The moving object color determining section 135 receives the correspondence information of moving objects and moving areas, moving area information A, moving area information B, and provisional moving object information from the moving body shape determining section 133, and performs the following steps. The representative chromaticity of the moving object corresponding to each moving object information is calculated according to the algorithm classified for each case. The representative chromaticity of the moving area information A and B is a vector that takes an integer value, but the representative chromaticity of the moving object information is a vector that takes a real value. Further, in each of the following equations (9) to (11), C _new is a new representative chromaticity of the moving body, C
_old is the estimated representative chromaticity of the moving object detected in the past, and α is a constant of 0 or more and 1 or less.

1) Moving area information A for moving objects detected in the past
If there is a corresponding moving area from both of the moving area information B and the moving area information B, the new representative chromaticity of the moving object detected in the past is determined according to (Equation 9).

[0190]

[Equation 9]

In (Equation 9), c _A is the representative chromaticity of the moving area corresponding to the moving object in the moving area information A, and c _B is the representative chromaticity of the moving area corresponding to the moving object in the moving area information B. .

2) Moving area information A for moving objects detected in the past
Alternatively, when only one of the moving area information B is associated with the moving area and the other is associated with the empty moving area, according to (Equation 10), a new representative of the moving object detected in the past is expressed. Determine chromaticity.

[0193]

[Equation 10]

In (Equation 10), c is the representative chromaticity of the non-empty moving area corresponding to the moving object.

3) The moving area information A and the moving area information B are added to the moving object information newly generated by the unsupported moving area processing unit 55.
If there is a corresponding motion area from both
According to 1), the representative chromaticity of the newly created moving body is determined.

[0196]

[Equation 11]

In (Equation 11), c _A is the moving area information A
Representative chromaticity dynamic area corresponding to the moving object in, c _B is a representative chromaticity dynamic area corresponding to the moving object in the moving region information B.

4) The moving area information A or the moving area information B is added to the moving object information newly generated by the unsupported moving area processing unit 55.
If there is a corresponding non-empty moving area from only one of the above, the representative chromaticity of the non-empty moving area is set as the representative chromaticity of the newly generated moving object.

According to the above algorithm, the moving body color determining unit 135 calculates the representative chromaticity of the moving body corresponding to each temporary moving body information, and outputs the temporary moving body information in which the representative chromaticity of each moving body is updated. To do.

The moving object information determining unit 136 updates the temporary moving object information in which the number of the moving object, the position of the moving object in the monitored area coordinates, the height of the moving object, and the representative chromaticity of the moving object are updated at the new current time. And outputs to the outside as the output of the distributed monitoring apparatus in this embodiment, and also to the moving body detection / evaluation circuit 61 of the moving body information storage means 4 and the moving body detection / evaluation means 5 in FIG. .

As described above, the moving body information calculation unit 58
Calculates the current moving object information, outputs the moving object information to the outside as an output of the distributed monitoring apparatus in the present embodiment, and detects moving object by the moving object information storage means 4 and the moving object detection evaluation means 5. Output to the evaluation circuit 61.

(Moving Body Information Storage Means) In FIG. 4, the moving body information storage means 4 receives the moving body information at the present time which is the output of the moving body information calculating section 58, stores it, and stores it until now. Of the past body object information stored, the body object information detected in the oldest frame is deleted. That is, the moving object information storage unit 4 sequentially stores the latest moving object information for the past N frames. In the present embodiment, the moving object information storage unit 4 stores moving object information corresponding to 30 frames, that is, one second of a moving image.

(Animal body detection / evaluation means) In FIG. 4, the animal body detection / evaluation circuit 61 of the animal body detection / evaluation means 5 stores the current animal body information and animal body information storage which are the outputs of the animal body information calculation section 58. Based on the past moving body information stored in the means 4, the current state of moving body detection is evaluated, and the moving body detection evaluation value is calculated according to the following algorithm. The animal body detection evaluation value in the present example indicates, as described above, the degree of whether the animal body is excessively detected as a result of the detection of the animal body, is underdetected, or is detected as an appropriate number. It is a value.

1) For one second immediately after the activation of the distributed monitoring apparatus in this embodiment, the moving object information storage means 4 does not store moving object information for 30 frames. In that case, the value indicating that a reasonable number has been detected is set as the moving object detection evaluation value, and the processing ends.

2) The error stored in the moving object information storage means 4
The average number of moving bodies is calculated from the previous moving body information. One
Mari, the physical condition of the past 30 frames stored
Calculate the number of animals in each frame of the report, 30 frames
Calculate the average in minutes and set it to N ₁far.

3) The number of moving bodies included in the moving body information at the present time, which is the output of the moving body information calculating unit 58, is calculated, and the calculated number is N ₂ .

4) If N ₂ is greater than N ₁ +1 then:
A value indicating that the animal body is excessively detected N ₂ −N ₁
-1 is set as the moving object detection evaluation value, and the process ends.

5) If N ₂ is smaller than N ₁ -1,
Value indicating that the body is under-detected N ₁ -N ₂
-1 is set as the moving object detection evaluation value, and the process ends.

6) In other cases, that is, N ₂ is N ₁ ± 1
When the value is within the range of, the value indicating that a reasonable number of moving objects has been detected is set as the moving object detection evaluation value.

According to the above algorithm, the moving object detection evaluation circuit 61 calculates the evaluation value of the current state of moving object detection. Then, the moving object detection evaluation value encoding circuit 62
Receives as input the moving object detection evaluation value output from the moving object detection evaluation circuit 61, encodes it so that it can be transmitted by the LAN 6 of FIG. 1, and outputs it via the communication means 35.

((Animal body information estimating means)) In FIG. 4, the animal body information estimating means 57 is a moving body information storing means 4
The current animal body information is estimated from the past animal body information stored in, and output as estimated animal body information. Hereinafter, the moving object information estimating means 57 will be described in more detail with reference to FIG.

FIG. 11 is a diagram showing the structure of the moving object information estimating means 57. In FIG. 11, reference numeral 141 denotes an animal body speed estimating means for estimating the speed of each animal body detected in the past from past animal body information, and 142 denotes each animal estimated by the past animal body information and the animal body speed estimating means 141. An estimated moving body position calculating means for estimating the current position of each moving body detected in the past from the estimated speed of the body, 143 is an estimated position of each moving body calculated by the estimated moving body position calculating means 142. Image coordinate calculation means 144 for obtaining the coordinates of each moving object in each image coordinate system of the moving image input means 1 and the moving image input means 7 is the current height of each moving object detected in the past from past moving object information. Body shape estimating means for estimating
Reference numeral 45 denotes an animal body color estimating means for estimating the present representative chromaticity of each animal body detected in the past from past animal body information, and the moving body speed estimating means 141 and the estimated moving body position calculating means 14
2 together constitute the moving body position estimating means 140. Not only the moving body speed estimating means 141, but also the moving body estimated position calculating means 142, the moving body shape estimating means 144, and the moving body color estimating means 145 are stored in the moving body information storage means 4. Referring to the information, FIG.
In, the reference line is omitted.

Next, the operation of the moving object information estimating means 57 will be described. The moving body speed estimation means 141 estimates the current speed of the moving body detected in the past from the past moving body information stored in the moving body information storage means 4 according to the following algorithm.

1) The following processing is performed in order from the moving object information having the smallest number in the moving object information storage means 4.

2) If there is a plurality of moving object information having the same number, the speed of the moving object corresponding to the number is estimated according to the following (Equation 12).

[0216]

[Equation 12]

In (Equation 12), V is the velocity of the animal body,
P _n is the position in the monitoring area coordinate of the moving object information of the newest frame in the moving object information having the same number, and P ₀ is the monitoring area of the moving object information in the oldest frame in the moving object information having the same number. Position in coordinates, f _n is the frame number of the newest frame in the body information having the same number, and f ₀ is the oldest frame in the body information having the same number. By number,
|| represents an absolute value.

3) When there is only one moving object information having a certain number, the speed of the moving object corresponding to that number is set to 0. That is, V in (Equation 12) is set to the 0 vector.

In this way, the animal body speed estimating means 14
1 estimates and outputs the speed of the moving object detected in the past.
Next, the moving object estimated position calculating means 142 calculates
According to 3), the current position of the moving object detected in the past is estimated.

[0220]

[Equation 13] In ( _Equation 13), P _predict is the estimated position of the _moving object,
f _c is the current frame number to be predicted,
f _n , V, and P _n are the same as those in (Equation 12), the frame number of the newest frame in the body information having the same number, the estimated speed of the body, and the body information having the same number. This is the position in the surveillance area coordinates of the moving object information of the newest frame in the. The moving body estimated position calculating means 141 calculates the estimated current position of the moving body detected in the past by (Equation 13). Then, the estimated moving body information is generated for each moving body detected in the past, and the estimated position of the moving body is set and output.

As described above, the moving body position estimating means 1
The reference numeral 40 estimates the current position of the moving object detected in the past from the moving object information in the past, and outputs the estimated moving object information in which the estimated position is set as one in the monitoring area coordinates.

The image coordinate calculating means 143 receives the estimated moving body information from the moving body position estimating means 140, and the moving image inputting means 1 from the estimated position in the monitored area coordinates of each moving body calculated by the moving body position estimating means 140. And the coordinates of the moving object in each image coordinate system of the moving image input means 7. The calculation is performed by the following (Equation 14) between the position in the surveillance area coordinate and the position in the image coordinate,
The fact that the relationship of (Equation 15) holds is used.

[0223]

[Equation 14]

[0224]

[Equation 15]

In (Equation 14) and (Equation 15), I _x ,
I _y is the image coordinates of a point on a plane captured by the camera, X and Y are the coordinates of a point on the plane in the coordinate system of that plane, a ₁
′, A ₂ ′, b ₁ ′, b ₂ ′, b ₃ ′, c ₁ ′, c ₂ ′, c ₃ ′ are related to the relationship between the camera and the plane, the plane coordinate system, and the image coordinate system of the camera. Dependent parameters, (Equation 1), (Equation 2)
Can be obtained by solving for I _x and I _y . If the moving object is out of the field of view of the camera, I _x ,
I _y may be a negative number or may exceed the number of effective pixels in the moving image, but the value as it is is used.

The image coordinate calculating means 143 calculates the current coordinates of the moving object detected in the past in the image coordinate systems of the moving image input means 1 and the moving image input means 7 as described above, and calculates the value thereof. The estimated moving object information is set and output.

The moving object shape estimating means 144 receives the estimated moving object information from the image coordinate calculating means 143, and displays the latest height of the moving object information corresponding to each number of each estimated moving object information as the moving object information. The value is read from the storage means 4, and the value is set as the estimated height of the moving body in the estimated moving body information and output.

The animal body color estimating means 145 receives the estimated animal body information from the animal body shape estimating means 144, and determines the latest representative chromaticity of the animal body information corresponding to each number of each estimated animal body information as an animal. Read from the body information storage means 4,
The value is set as the estimated representative chromaticity of the moving body in the estimated moving body information and output.

As described above, the moving body information estimating means 5
Reference numeral 7 estimates the current position of each moving body from the moving body information of the moving body detected in the past stored in the moving body information storage unit 4, and outputs it as estimated moving body information.

As described above, the moving object detecting means 3 in this embodiment uses the moving area information input from the moving image inputting means 1 and the moving image inputting means 7 from the moving body information of the moving body detected in the past. By using the estimated moving object information calculated by the moving object information estimating unit 57 and associating it with the previously detected moving object, the moving object can be detected and the moving object information can be calculated.

(Effect of Embodiment 1) As described above,
The distributed monitoring apparatus according to the first embodiment processes moving images captured by a plurality of moving image inputting means 1 and 7 by moving image processing means 2 and 8, and obtains a moving region as moving object detecting means. By integrating in 3 in the above, even if the moving image cannot be detected in the moving image input from one moving image input means due to a failure or concealment, the other camera can complement the moving image. It is possible to perform monitoring with a small error in detecting a moving object. For example, by arranging the television camera of the moving image input means 1 and the television camera of the moving image input means 7 so that the shooting directions are at right angles, the frame will look like when the moving object is moving toward one of the cameras. Even if detection of the moving area fails because the change in the area corresponding to each moving object is small and the difference cannot be detected well, the other camera easily detects the difference because it moves vertically to the main axis of the camera. Therefore, the detection error of the moving object can be prevented. Also, even if the moving areas corresponding to different moving objects are fused in one TV camera due to the passing of the moving objects, they may be separated in the other TV camera. Can be prevented.

Further, by storing the moving body information of the moving body detected in the past in the moving body information storage means 4 and integrating the information, even if the detection of the moving body is temporarily failed, it is stored. It is possible to compensate for this, and it is possible to perform monitoring with less error in detecting a moving object. For example, even if the moving regions corresponding to different moving objects are temporarily fused due to passing or the like, if the moving objects are correctly detected before the passing, the moving objects can be correctly detected even during the passing.

Furthermore, the moving images output from the moving image input means 1 and the moving image input means 7, the moving area information output from the moving image processing means 2 and the moving image processing means 8, and the moving object output from the moving object detecting means 3. Since none of the information is information that directly depends on the internal processing, the moving image processing method of the moving image processing means 2 and the moving image processing means 8 is different from that of the present embodiment, and different methods are used respectively. However, the influence on other parts can be minimized. Even if only the moving image processing means 8 is changed in the middle, the influence on other parts can be minimized. Therefore, when the algorithm or the hardware is changed, other parts are less affected and the monitoring device is highly reliable.

Further, the position of the moving object corresponding to the moving area calculated by the area position calculating means 28 of the moving image processing means 2 shown in FIG. 3 in the coordinates of the monitoring area is calculated, and the position information thereof is detected by the moving object detecting means. By using it in association with the moving object detected in 3 in the past, it is possible to improve the reliability of the integration of the plurality of moving image input means and the integration of the past moving object detection information.

The area position calculating means 28 constituted by the plane contact point coordinate calculating means 94 and the space coordinate calculating means 93.
Calculate the position in the surveillance area coordinates of each moving area by
By using the position information for associating the moving area with the moving object detected in the past and using the position information for integration in the moving object detection means, it is possible to detect the moving object more accurately.

Further, the moving object information estimating means 57 estimates the moving object information of the moving object detected in the past, and outputs the estimated moving object information and the current moving image processing means 2 and moving image processing means 8. By associating the moving body area information with the moving body body moving area associating unit 51, the reliability of the integration of the plurality of moving image inputting units and the integration of the past moving object detection information can be improved.

Further, the current position of the moving object previously detected by the moving object position estimating means 140 is estimated, and the estimated current position is used for the moving object moving area associating means 51 to associate the moving object with the moving area. This makes it possible to associate a moving object with a moving object with higher reliability, and as a result, perform highly reliable monitoring.

The area color calculating means 102 calculates the color of the moving area, and the moving object color estimating means 145 estimates the color of the moving object detected in the past.
By using it for associating the moving object with the moving object in 1 above, the moving object can be associated with the moving object with higher reliability, and as a result, highly reliable monitoring can be performed.

The area shape calculating means 103 calculates the height of the moving body corresponding to the moving area, and the moving body shape estimating means 145 estimates the height of the moving body detected in the past to calculate the moving body moving area. By using the association of the moving object and the moving object in the associating unit 51, the moving object and the moving object can be associated with higher reliability, and as a result, highly reliable monitoring can be performed.

Further, the moving object position estimating means 140 estimates the current position of the moving object detected in the past, and the concealing object position storing means 121 stores the position of the concealing object existing in the surveillance area. If it is determined that the moving object is hidden by the hiding object by determining from the estimated current position of the moving object and the position of the hiding object whether or not the previously detected moving object is hiding by the hiding object By determining that there is no moving area to be moved, it is possible to prevent an error in associating the moving object with the moving body when the corresponding moving area cannot be detected by one moving image input means.

Further, the current position of the moving body detected in the past is estimated by the moving body position estimating means 140, and the moving body detected in the past is hidden by the moving body hidden by the moving body concealment judging means 124. By determining from the estimated current position of the moving object, it is possible to prevent an error in associating the moving object with the moving area when the corresponding moving area cannot be detected by one moving image input means.

Further, the coordinate values of the moving image of the moving image input means 1 and the moving image input means 7 corresponding to the estimated current position of the moving object detected in the past by the image coordinate calculating means 143 are obtained in the image coordinate system, Shooting range determination means 125 using
When it is determined whether or not the moving object detected in the past is imaged by the moving image input means 1 and the moving image input means 7, and the moving object detected in the past is outside the imaging range of the moving image input means. It prevents the mistaken correspondence between the moving body and the moving area.

Further, the present position of the moving object detected in the past is estimated by the moving object position estimating means 140, and the moving image inputting means position storing means 123 stores the moving image inputting means 1 and the moving image inputting means 7 in the monitoring area coordinates. If the distance between the estimated current position of the moving object and each moving image input means exceeds the distance threshold stored in the distance threshold storage means 126, the moving body detected in the past is input. By determining that the moving object cannot be associated with the moving area because it is too far from the means, the moving object is detected by the moving image input means because the moving object detected in the past is within the imaging range of the moving image input means but too far. Even if it is not possible, it is possible to prevent an error in associating the moving object with the moving object.

Further, the moving body detection evaluation means 5 evaluates the result of the moving body detection in the moving body detection means 3, and the diaphragm of the television camera 11 is adjusted by the moving image input adjusting means 13 according to the evaluation result. In addition, the diaphragm of the TV camera can be kept optimal for detecting a moving object even when the lighting conditions change, and the time and effort for adjustment at the time of installation can be saved.

Further, the moving object detection evaluating means 5 evaluates the result of the moving object detection in the moving object detecting means 3, and the threshold value of the threshold value processing circuit 25 is adjusted through the moving image processing adjusting means 30 according to the evaluation result. As a result, it is possible to keep the state of the moving image processing in a better state and save the labor of calculating and setting the threshold value at the time of installation.

Further, the moving image processing evaluation means 72 evaluates the result of the moving image processing in the moving image processing means 2, and the threshold value of the threshold value processing circuit 25 is adjusted via the moving image processing adjustment means 30 according to the evaluation result. As a result, the state of moving image processing can be maintained in a better state, and the labor for calculating and setting the threshold at the time of installation can be saved.

(Second Embodiment) The second embodiment of the present invention will be described below with reference to the drawings. FIG. 12 is a diagram showing the configuration of the distributed monitoring apparatus according to the second embodiment of the present invention. In FIG. 12, 201 and 204 are moving image input means for picking up a moving image of the surveillance area and outputting it as a monochrome digital video signal, and 202 and 205 are monochrome digital video signals output by the moving image input means 201 and 204, respectively. Moving image processing means for processing and outputting moving area information, 203 and 206 are moving image processing means 202, respectively.
And 204 to receive an output of the moving object, detect the moving object, calculate the moving object information, and output the moving object to the outside, 20
7 and 208 are the moving body detecting means 203 and the moving body detecting means 2
The moving body information storage means for storing the moving body information which is the output of 06, 209 is a LAN for communicating between the moving body detecting means 203 and 206, and 241 and 242 are outputs from the moving image processing means 202 and 205, respectively. It is a moving image processing evaluation means for calculating an evaluation value for calculation of certain moving area information. Hereinafter, moving image input means 201 and 204 and moving image processing means 202
And 205 and the moving image processing evaluation means 241 and 242, the operation of the moving object detection means 203 and 206 and the moving object information storage means 207 and 208 will be described.

(Moving Image Input Means) First, the moving image input means 201 will be described with reference to FIG. Since the moving image input means 204 has the same configuration and operation as the moving image input means 201, duplicate description will be omitted.

FIG. 13 is a diagram showing the structure of the moving image input means 201. In FIG. 13, reference numeral 211 denotes a television camera having a function of adjusting an aperture from the outside, which captures a moving image and outputs it as a monochrome analog video signal, and 212 denotes A which converts the monochrome analog video signal output from the television camera into a digital signal. It is a / D converter. Reference numeral 213 is a moving image input adjusting means for changing the state of moving image input, and an aperture adjusting device 214 for adjusting the aperture of the television camera 211.
And a diaphragm adjustment amount calculation circuit 215 for calculating the diaphragm adjustment amount of the diaphragm adjustment device 214 from the moving image processing evaluation value output from the moving image processing evaluation means 241.

Next, the operation of the moving image input means 201 will be described. In the present embodiment, the television camera 211 of the moving image input unit 201 and the television camera of the moving image input unit 204 image areas that do not overlap each other. The monochrome analog video signal captured by the TV camera 211 is A /
The digital signal is converted by the D converter 212 and output to the moving image processing means 202 in FIG. On the other hand, the aperture adjustment amount calculation circuit 215 receives the moving image processing evaluation value from the moving image processing evaluation means 241 in FIG. In the present embodiment, the moving image processing evaluation value is a value which is 0 or a positive value, which is an evaluation value of the moving image processing state of the moving image processing means 202. The aperture adjustment amount calculation circuit 215 determines the aperture adjustment amount by the algorithm classified according to the following cases and outputs it to the aperture adjustment device 214.

1) If the moving image processing evaluation value is positive and the aperture adjustment has not been performed in the near past, an adjustment amount is randomly generated, and the television camera 21 via the aperture adjustment device 214.
Adjust the aperture of 1.

2) When the moving image processing evaluation value is positive and the aperture adjustment is performed in the near past, the aperture adjustment performed in the near past is compared with the past moving image processing evaluation value, and the moving image is moved more than the past. If the processing evaluation value is small, or if there is no change, the same aperture adjustment amount as in the past is output, and if the moving image processing evaluation value is larger, the aperture adjustment that is the opposite of the positive and negative values of the past is output. The amount is output and the aperture of the television camera 211 is adjusted via the aperture adjusting device 214.

3) When the moving image processing evaluation value is 0, the aperture adjustment amount calculation circuit 215 outputs 0 as the aperture adjustment amount and does not perform aperture adjustment.

Then, the aperture adjustment device 214 adjusts the aperture of the television camera 211 according to the aperture adjustment amount output from the aperture adjustment amount calculation circuit 215.

As described above, the moving image inputting means 201 in this embodiment picks up the image of the monitoring area, outputs the result as a monochrome digital image signal, and uses the moving image processing evaluation value to improve the aperture. By doing so, it is possible to output a moving image in which the moving region is appropriately detected even if the illumination condition of the monitoring area changes.

(Moving Image Processing Means) Next, the moving image processing means 202 of FIG. 12 will be described with reference to FIG. The moving image processing means 205 and the moving image processing evaluation means 2
Since 42 has the same configuration and operation as the moving image processing means 202 and the moving image processing evaluation means 241, respectively, duplicate description will be omitted.

FIG. 14 is a diagram showing the structure of the moving image processing means 202 together with the moving image processing evaluation means 241. Figure 1
4, a frame buffer 222 stores a frame image of one frame of the monochrome digital video signal output from the A / D converter 212 of FIG. 13, and a frame buffer 223 stores the frame image stored in the frame buffer 222. While the next frame image is stored in
A frame buffer 224 that is temporarily stored, a difference calculation circuit that generates a difference image of the brightness value of each pixel of the frame images stored in the frame buffer 222 and the frame buffer 223, and 225 is a difference image output from the difference calculation circuit 224. Of the respective pixels of, the threshold value processing circuit 226 for generating a binarized difference image in which the pixel having a value exceeding the predetermined threshold value is 1 and the other pixels are 0, An area shaping unit 227 for shaping the connected area of the binarized difference image, a labeling processing circuit 227 for labeling the connected area in the binarized difference image output from the area shaping unit 226 to generate a labeling image, 228.
Is an area position calculation means for calculating the position of each label area of the labeling image, which is the output of the labeling processing circuit 227, in the monitoring area coordinates, and 229 is the labeling processing circuit 22.
The shape feature amount of each label region of the labeling image which is the output of No. 7 is calculated, and the feature amount and the region position calculating means 22 are calculated.
8 is a moving area information calculation unit that collectively outputs the position of each label area, which is the output of 8, as moving area information. Reference numeral 241 denotes a moving image processing evaluation unit that evaluates the calculation result of the moving region information output from the moving region information calculation unit 229 and outputs a moving region processing evaluation value, and a portion excluding the moving image processing evaluation unit 241. Constitutes the moving image processing means 202.

Next, the operation of the moving image processing means 202 will be described. The frame buffer 222 receives a moving image as a monochrome digital video signal from the A / D converter 212 in FIG. 13 and stores an image for one frame. The image for one frame stored in the frame buffer 222 is copied to the frame buffer 223 before the image for the next one frame is stored in the frame buffer 222. The difference calculation circuit 224 obtains the difference between the pixel values of successive frame images stored in the frame buffer 222 and the frame buffer 223 and generates a difference image. The threshold value processing circuit 225 sets the pixel value to 1 when each pixel value of the difference image generated by the difference calculation circuit 224 exceeds a predetermined threshold value, and sets the pixel value to 0 otherwise, and is a binary difference image. To generate. Next, the area shaping unit 226 shapes the connected area of the binarized difference image generated by the difference calculation circuit to generate a binarized difference image. The area shaping unit 226
Is the area shaping unit 26 in the first embodiment shown in FIG.
Since the configuration and the operation are the same, detailed description thereof will be omitted.

The binarized difference image subjected to the area shaping processing by the area shaping unit 226 is input to the labeling processing circuit 227 and subjected to the labeling processing. The labeling process in the present embodiment assigns a number to each connected region included in the input binary difference image, and generates a labeling image in which the number of the connected region is set as the pixel value of the pixel of each connected region. To do.

The area position calculation means 228 calculates the position of the moving area corresponding to each label in the coordinates of the monitoring area from the labeling image output from the labeling processing circuit 227. This position can be regarded as the position of the moving body corresponding to each moving area. Since the configuration and operation of the area position calculating means 228 are the same as those of the area position calculating means 28 in the first embodiment shown in FIG. 6, detailed description thereof will be omitted.

In the moving area information calculation unit 229, the labeling image output from the labeling processing circuit 227 and the position of the moving object corresponding to each moving area output from the area position calculating unit 228 in the monitoring area coordinates are input, The moving area information including the moving area corresponding to the moving object in the monitoring area and its feature amount is output. In the present embodiment, the moving area information is the label number of each moving area, the image coordinates of the upper left and lower right of the circumscribing rectangle of each moving area, and the position in the monitoring area coordinates of the moving object corresponding to each moving area, It is composed of the height of the corresponding moving body. Hereinafter, with reference to FIG. 16, the moving area information calculation unit 2
29 will be described in more detail.

FIG. 16 is a diagram showing the configuration of the moving area information calculation unit 229 together with related circuits. In FIG. 16, reference numeral 261 denotes a circumscribed rectangle calculation circuit that calculates the circumscribed rectangle of each moving area of the labeling image output from the labeling processing circuit 227 and calculates the image coordinates of the upper left and lower right thereof, and 262 denotes the circumscribed rectangle calculation circuit 261. From the calculated image coordinates of the vertices of the circumscribing rectangle and the position of the moving object corresponding to each moving area in the monitored area coordinates, which is the output of the area position calculating means 228, the height of the moving object corresponding to each moving area is calculated. It is a region shape calculation means for calculating.

Next, the operation of the moving area information calculation section 229 will be described. The circumscribing rectangle calculation circuit 261 obtains a circumscribing rectangle in which each side is parallel or perpendicular to the axis of the image coordinates of each moving area of the labeling image output from the labeling processing circuit 227, and determines the upper left vertex and the lower right vertex. The image coordinate value is output together with the label number of each moving area. Area shape calculation means 2
Reference numeral 62 indicates the height of the circumscribing rectangle in the image from the coordinate values of the vertices of the circumscribing rectangle output from the circumscribing rectangle calculating circuit 261, and the height of the circumscribing rectangle is the output of the area position calculating means. The height of the moving object corresponding to each moving area is calculated from the position of the moving object in the monitored area coordinates, and the focal length and position of each camera stored in advance.
Then, the label number of each moving area, the image coordinate value of the circumscribing rectangle, and the position and height of the corresponding moving object in the monitored area coordinates are output as moving area information.

As described above, the moving area information calculation unit 22
Reference numeral 9 denotes the labeling processing circuit 227 and the area position calculating means 2
It receives an input from 28 and outputs moving area information.

As described above, the moving image processing means 202 in this embodiment can calculate and output moving area information from the monochrome digital video signal input from the moving image input means 201.

(Moving image evaluation processing means) In FIG.
The moving image processing evaluation means 241 receives the moving area information from the moving image processing means 202, and if the height of the moving object corresponding to each moving area is smaller than a preset threshold value, the difference from the threshold value is calculated. In other cases, 0 is set as the moving region processing evaluation value, and the result is output to the aperture adjustment amount calculation circuit 215 in FIG. The height threshold set in advance here is the lower limit of the typical height of the moving object to be monitored in the monitoring area, and the height of the moving object corresponding to each moving area is higher than the threshold. The small size means that the moving image input by the moving image input unit 201 has a poor contrast and the moving region is detected smaller than the actual size.

(Moving Object Detecting Means) Next, the moving object detecting means 203 will be described with reference to FIG. In addition,
Since the configuration and operation of the moving object detecting unit 206 are the same as those of the moving object detecting unit 203, duplicate description will be omitted.

FIG. 15 is a diagram showing the structure of the moving body detecting means 203 in this embodiment together with the moving body information storage means 207. 15, reference numeral 252 indicates a moving area and a moving object detected in the past based on the moving area information output from the moving image processing means 202 of FIG. 12 and the estimated moving object information output from the moving object information estimating means 259. The feature association unit 254 that associates the feature amount with the feature amount estimates the reason why the moving object detected by the feature association unit 252 was not associated with the moving object that was detected in the past, and performs the corresponding process. The unsupported moving object processing unit 255 estimates the reason why the moving region that has not been matched with the moving object detected in the past by the feature matching unit 252 is not matched, and performs the corresponding process. The area processing unit 256 outputs the unsupported moving object processing unit 254 and the unsupported moving area processing unit 255.
Is a correspondence determining unit that determines the correspondence between the moving region and the moving object detected in the past, and these configure the moving body moving region matching unit 251. 257 is a moving object information that detects a moving object existing in the monitored area from the matching result of the moving area output from the matching determining unit 256 and the moving object detected in the past, and calculates and outputs the moving object information. The calculation unit 258 encodes the moving object information output from the moving object information calculating unit 257 so that it can be transmitted by the LAN 209, and the reference numeral 259 indicates the past moving object stored in the moving object information storage unit 207. Animal body information estimating means for estimating the current animal body information of the previously detected animal body from the information, and 260 is an animal body information decoding circuit for decoding the encoded animal body information transmitted by the LAN 209. Thus, the moving object detection unit 203 is configured by these and the moving object moving region associating unit 251.

Next, the operation of the moving object detecting means 203 will be described. The feature associating unit 252 of the moving object moving region associating unit 251 receives the moving region information which is the output from the moving region information calculating unit 229 of FIG. 14, and detects the moving body which is the output of the moving body information estimating unit 259. The means 203 and the moving body detection means 206 receive the estimated moving body information of the moving body detected in the past, and associate the moving region with the moving body detected in the past. Hereinafter, each part of the animal body motion area associating means 251 will be described in more detail.

((Animal body moving area associating means)) ((feature associating portion)) FIG. 17 is a diagram showing the structure of the feature associating portion 252 in the animal body moving area associating means 251. In FIG. 17, reference numeral 271 denotes a distance association circuit that receives inputs from the moving area information calculation unit 229 and the moving object information estimating unit 259 and associates the moving area with the moving object detected in the past based on the distance in the monitored area coordinates. 27
2 receives an input from the distance association circuit 271, and uses a shape feature amount to associate a moving region with a previously detected moving object, and a shape association circuit 273 receives an input from the shape association circuit 272. Is a correspondence determination circuit that finally determines correspondence between a moving area and a moving object detected in the past.

In the present embodiment, the moving area information corresponds to each moving area, the label number of each moving area, the upper left and lower right image coordinates of the circumscribed rectangle of each moving area, as described above. It is composed of the position and height of the moving object in the surveillance area coordinates. The moving body information includes the moving body number, the position and height of the moving body in the monitored area coordinates, and the moving image input means 201.
And the position of the moving object in the image coordinate system of the moving image captured in.

Further, the feature associating unit 2 in this embodiment.
In FIG. 52, the correspondence between the moving area information and the moving body information is shown in FIG.
The moving area output from the second moving image processing unit 202 is associated with the moving object detected in the past. In addition, the moving object detection means 2
The feature association unit 06 associates the moving area output from the moving image processing unit 205 with the moving object detected in the past.

The operation of the feature associating unit 252 will be described below. The distance association circuit 271 includes moving area information obtained as a result of processing the moving image captured by the moving image input unit 201 from the moving area information calculating unit 229 in FIG. 14, and the moving object information estimating unit 259 in FIG. And the estimated moving object information of the moving object detected in the past. Then, the association is performed in the following steps.

1) Select the estimated moving object information corresponding to the moving object that has not been subjected to the association process among the moving objects detected in the past.

2) The Euclidean distance between the estimated position of the currently selected moving object in the monitored area and the position of the moving object to which each moving region corresponds is calculated.

3) Of the moving areas, those having the Euclidean distance calculated in step 2 larger than a preset threshold value are removed from the candidates for association with the currently selected moving object.

4) The numbers of the remaining moving areas are arranged in the ascending order of the Euclidean distance calculated in step 2, and are set as the association candidates for the currently selected moving object.

5) If there is a moving object that has not been associated, the process returns to step 1; otherwise, the process ends.

Through the processing by the above steps, the moving area is associated with each moving object previously detected by the distance associating circuit 271. However, one moving body does not always correspond to one moving area, and there may be a case where there is no corresponding moving area or a plurality of corresponding moving areas. Then, the distance correspondence circuit 2
71 outputs the association candidate information based on the distance together with the input moving area information and the estimated moving object information of the moving object detected in the past.

Next, the shape associating circuit 272 outputs, as an output of the distance associating circuit 271, candidate information for associating a moving area with a previously detected moving object, moving area information, and an estimated moving object of a moving object detected in the past. Input the information and perform the following steps.

1) The following steps 2 and 3 are performed for each moving object detected in the past.

2) The difference between the height of each moving region of the currently selected moving object association object and the estimated height of the moving object is calculated.

3) The shape correspondence evaluation value is set to 0 for the moving area of the association candidate whose difference calculated in step 2 exceeds the preset threshold value, and the moving area of the association candidate below the preset threshold value is the shape. The corresponding evaluation value is 1.

As described above, the shape correspondence circuit 27
2 calculates the shape correspondence evaluation value by using the height of the moving body detected in the past and the height of the moving body corresponding to the moving area as shape information, and using the difference as an index of similarity. Then, together with the input moving area information and the estimated moving object information of the moving object detected in the past, the association candidate information based on the distance with the shape evaluation value added is output.

Next, the association determining circuit 273 receives the moving area information, the estimated moving object information of the moving object detected in the past, and the association candidate information to which the shape evaluation value is added from the shape associating circuit 272. In the following steps, at most one moving region corresponding to each moving object detected in the past is determined.

1) The following steps 2 to 4 are performed for each moving object detected in the past.

2) The shape correspondence evaluation values of the moving area correspondence candidates are compared, and if there is one moving area, it is determined that the moving area corresponds to the moving object. If there is a plurality, it is determined that the first moving area in the order arranged by the distance correspondence circuit 271 corresponds to the moving object.

As described above, the feature associating unit 252
Determines at most one moving region corresponding to the moving object detected in the past, considering the distance and shape in the coordinate system in the real space of the monitored area. Then, together with this association result, the input motion area information and the estimated moving object information of the moving object detected in the past are output.

((Unsupported moving object processing unit)) The unsupported moving object processing unit 254 in FIG. 15 associates the moving area output from the feature associating unit 252 with the moving object detected in the past,
Processing is performed when a moving area is not associated with a moving object detected in the past. Hereinafter, the unsupported moving object processing unit 252 will be described in more detail with reference to FIG.

FIG. 18 is a diagram showing the structure of the unsupported moving object processing unit 254. In FIG. 18, reference numeral 281 denotes a moving object hiding determination unit that determines whether a previously detected moving object is hidden by another previously detected moving object, and 282 indicates a position of the moving image input unit 201 in the monitoring area coordinates. The moving image input unit position storing unit 283 stores the moving image input unit position storing unit 283. The operation of the unsupported moving object processing unit 254 is as follows.

The moving image input means position storage means 282 stores
The position of the moving image input means 201 in the surveillance area coordinates is stored in advance. The object concealment determination means 281 is
The correspondence result of the moving area and the moving object detected in the past, the moving area information, and the estimated moving object information of the moving object detected in the past are received from the feature matching unit 252, and the processing is performed in the following steps.

1) Of the moving objects detected in the past, one having no corresponding moving area is selected, and the process of step 2 is performed.

2) A moving image inputting means position storing means 282 calculates a line segment connecting the position of the moving image inputting means 201 and the position of the selected moving object in the monitoring area coordinates, and calculates the movement of the line segment. From the area information, the distance to the estimated position of the moving object that has the corresponding moving area is calculated. Then, if there is a moving object that has a corresponding moving area such that the distance is equal to or less than a preset threshold value, it is determined that the selected moving object is hidden by the moving object,
An empty moving area is set as the corresponding moving area.

In this way, the moving object hiding judging means 28
1 correctly determines the reason why the object cannot be associated when the object is hidden by another object. Then, the result of associating the moving area with the moving object detected in the past, the moving area information, and the estimated moving object information of the moving object detected in the past are output.

Next, the photographing range judging means 283 receives from the moving object hiding judging means 281 the result of associating the moving area and the moving object detected in the past, the moving area information and the estimated moving object information of the moving object detected in the past. , The process is performed in the following steps.

1) Of the moving objects detected in the past, one having no corresponding moving area in the moving area information is selected, and the process of step 2 is performed.

2) When the coordinates of the moving object in the image coordinate system of the moving image picked up by the moving image input means 201 of the selected moving object information is outside the image range of the moving image of the moving image input means 201. , The selected moving object is moving image input means 2
It is determined that the moving area is outside the imaging range of 01, and an empty moving area is set as the corresponding moving area.

As described above, the photographing range judging means 283 is
Using the coordinates of the moving object in the image coordinate system of the moving image input means 201, the reason why the moving object cannot be associated is correctly determined when the moving object is outside the imaging range of the television camera.

As described above, the unsupported moving object processing unit 2
Reference numeral 54 denotes a moving object detected in the past when the moving object detected in the past is hidden by another moving object and when the moving object is outside the moving image capturing range of the moving image input unit 201. The reason why the association cannot be performed correctly is determined, and when the reason is given, the association result in which the empty moving area is set as the corresponding moving area is added to the association information output from the feature associating unit 252, and The region information and the estimated moving object information of the moving object detected in the past are output together.

((Non-corresponding motion area processing unit)) The non-corresponding motion area processing unit 255 of FIG. 15 newly adds to the motion area not previously associated with the moving object detected by the feature association unit 252. The moving body information is generated, and each moving area is associated with each moving area corresponding to the moving body. Then, the newly generated moving object information and the correspondence information between the moving object information and the moving area are output.

((Association Confirmation Unit)) Next, the association determination unit 256 determines the moving area information, the estimated moving object information of the previously detected moving object and the previously detected moving object from the unsupported moving object processing unit 254. And the moving area association result are input, and the moving object processing unit 255 newly creates moving object information from moving areas that have not been associated with moving objects previously detected by the feature associating unit 252. And the correspondence information of the moving area corresponding thereto are input. Then, by the processing of the following steps, the association between the moving object detected in the past and the moving area is determined and output as moving object information.

1) The correspondence information of the moving body and the moving area input from the unsupported moving body processing unit 254 and the correspondence information of the moving body and the moving area input from the unsupported moving body processing unit 255 are combined. It collects in one correspondence information.

2) Of the moving object information, the moving object information having no corresponding moving area is removed.

Through the above processing, the association determining unit 256
Determines the correspondence between the moving object and the moving object detected in the past, and the matching information, moving area information, estimated moving object information of the moving object detected in the past, and newly generated moving object information. And output.

As described above, the moving body moving area associating means 251 makes the moving body information detected in the past and the moving area information calculated by the moving image processing means 202 correspond to each other, and the moving body moving objects which are not made correspond to each other. For that, if the reason is not attached, the moving object is deleted, and a moving object that is not associated with the moving object is newly generated and associated, and the moving object and the moving area are associated with each other. Associated information, motion area information,
The estimated moving body information of the moving body detected in the past and the newly generated moving body information are output.

((Animal body information calculation unit)) The animal body information calculation unit 257 of FIG.
The moving body information storage means that receives the output information of the correspondence between the moving body and the moving area, the moving area information, the estimated moving body information of the previously detected moving body, and the newly generated moving body information. The current moving object information is calculated with reference to the past moving object information stored in 207. Then, the moving object information is output to the outside as an output of the distributed monitoring apparatus according to the present embodiment, and is also output to the moving object information storage unit 207 and the moving object information encoding circuit 258. It should be noted that, of the moving object information, the position of the moving object in the image coordinate system of the moving image inputting means 201 is not required as an output of the moving object information calculation unit in the present embodiment, and therefore no new calculation is performed. Hereinafter, the moving object information calculation unit 257 will be described in more detail with reference to FIG.

FIG. 19 is a diagram showing the structure of the moving object information calculation unit 257. In FIG. 19, 291 is a moving body number determining unit that determines the number of each moving body, 292 is a moving body position determining unit that calculates the position of each moving body in the surveillance area coordinates, 29
Reference numeral 3 denotes a moving body shape determining unit that calculates the height of each moving body. The operation of the moving object information calculation unit 257 will be described below.

[0308] The moving body number determining unit 291 outputs the moving body moving area associating means 251 associating information between moving bodies and moving areas, moving area information, and estimated moving body information of previously detected moving bodies. And the newly generated moving object information are input. Then, the number of the estimated moving object information of the previously detected moving object output from the unsupported moving object processing unit 254 is not changed while referring to the previous moving object information stored in the moving object information storage unit 207. , Body information storage means 207
The numbers other than the numbers used in the past moving object information stored in are assigned to the newly generated moving object information in ascending order. Furthermore, the estimated moving body information of the moving body detected in the past and the newly generated moving body information are combined into one as provisional moving body information, which is output together with the correspondence information of the moving body and the moving area and the moving area information. To do. Note that the numbers do not have to be consecutive after the numbers are assigned because the number of numbers removed by the association determining unit 256 is larger than the number of newly generated moving object information.

Next, the moving object position determining unit 292 receives the correspondence information between the moving objects and the moving area, the moving area information and the provisional moving object information numbered from the moving object number determining unit 291, and the following is performed. The position of the moving body in the monitoring area coordinates corresponding to each moving body information is estimated according to the algorithm classified for each case.

1) When a moving region that is not empty corresponds to a moving object detected in the past, a new position in the monitored area of the moving object detected in the past is determined according to (Equation 16).

[0311]

[Equation 16]

In (Equation 16), P _new is the new position of the moving object detected in the past, P _old is the estimated position of the moving object detected in the monitoring area coordinates in the past, and p is the monitoring area coordinates of the corresponding moving area. The position and α are constants of 0 or more and 1 or less.

2) When the moving body in the sky corresponds to the moving body detected in the past, the estimated position of the moving body detected in the past in the monitoring area coordinates is set as the new position in the monitoring area of the moving body detected in the past. To do.

3) If the moving object information is new moving object information generated by the unsupported moving area processing unit 255, the position of the moving area in the monitored area coordinates is set as the position of the moving object.

In accordance with the above algorithm, the moving body position determining unit 292 calculates the position of each provisional moving body information in the coordinates of the moving body monitoring area, and the correspondence information between the moving body and the moving area, the moving area information, and each animal. The provisional moving object information in which the position of the body in the monitored area coordinates is updated is output.

Next, the moving body shape determining unit 293 receives the moving body area / moving area correspondence information, the moving area information, and the provisional moving body information from the moving body position determining unit 292, and classifies them in the following cases. The height of each moving object information in the monitoring area coordinates of the moving object is calculated according to the algorithm.

1) When there is a non-empty moving area corresponding to the moving object detected in the past, the new height of the moving object detected in the past is determined according to (Equation 17).

[0318]

[Equation 17]

In (Equation 17), H _new is the new height of the moving body detected in the past, H _old is the estimated height of the moving body detected in the past, h is the height of the corresponding moving region, and α is 0. The constant is 1 or less.

2) When the moving object in the sky corresponds to the moving object detected in the past, the estimated height of the moving object detected in the past in the monitoring area coordinates is set to the new height in the monitoring area of the moving object detected in the past. Satoshi

3) When the moving object information is new moving object information generated by the unsupported moving area processing unit 255, the height of the moving area is set as the height of the moving object.

In accordance with the above algorithm, the moving object shape determining unit 293 calculates the height of each moving object in the provisional moving object information. Then, the number of the moving object, the position of the moving object in the monitoring area coordinates, and the provisional moving object information in which the height of the moving object is updated are set as the new moving object information at the present time and output as the output of the monitoring device in the present embodiment. At the same time, the data is output to the moving object information storage unit 207 and the moving object information encoding circuit 258 in FIG.

As described above, the moving body information calculation unit 25
Reference numeral 7 calculates the current moving object information of the moving object detected within the imaging range of the moving image inputting means 201, and outputs the moving object information to the outside as an output of the distributed monitoring apparatus according to the present embodiment. The data is output to the body information storage unit 207 and the moving body information encoding circuit 258.

The moving object information encoding circuit 258 outputs the moving object information at the present time, which is the output of the moving object information calculation section 257, to the L level.
LAN209 is encoded so that it can be transmitted by AN209.
It is sent to the moving object detecting means 206 via. The moving object information decoding circuit 260 receives the current moving object information detected by the moving object detecting unit 206 from the moving object detecting unit 206 via the LAN 209, decodes it, and outputs it to the moving object information storage unit 207.

(Moving Body Information Storage Means) Next, the moving body information storage means 207 will be described. Since the moving object information storage unit 208 has the same configuration and operation as the moving object information storage unit 207, duplicate description will be omitted.

The moving object information storage unit 207 receives from the moving object information calculation unit 257 the moving object information detected within the current image pickup range of the moving image input unit 201, and
The moving object detecting means 2 via the moving object information encoding circuit 260.
The moving object information within the imaging range of the moving image input means 204 at the present time detected in 06 is received. As mentioned above,
Since the image capturing ranges of the moving image input means 201 and the moving image input means 204 do not overlap, the moving object information from both corresponds to different moving objects.

The moving object information storage means 207 stores these moving object information and deletes the moving object information detected in the oldest frame among the past moving object information stored so far. That is, the latest moving object information for the past N frames is sequentially stored. In the present embodiment, the moving object information storage unit 207 stores moving object information corresponding to 90 frames, that is, 3 seconds of the moving image.

((Motor Body Information Estimating Means) Next, referring to FIG.
The moving object information estimating means 259 will be described in more detail with reference to FIG. FIG. 20 is a diagram showing the structure of the moving object information estimating means 259. In FIG. 20, 3
Reference numeral 01 is an animal body speed estimating means for estimating the speed of each animal body detected in the past from past animal body information, and 302 is the speed of each animal body estimated by the past animal body information and the animal body speed estimating means 301. From the estimated position of each moving object that is estimated by the moving object estimated position calculating means 302, a moving image is input from the estimated moving object position calculated by the estimated moving object position calculating means 302. Image coordinate calculation means for calculating the image coordinates of each moving object in the image coordinate system of the means 201, 304 is a moving object shape estimating means for estimating the current height of each previously detected moving object from past moving object information. Therefore, the moving object position estimating means 300 and the moving object estimated position calculating means 302 are included in the moving object position estimating means 300.
Are configured. Note that not only the moving body speed estimating means 301, but also the moving body estimated position calculating means 302 and the moving body shape estimating means 304 refer to the moving body information stored in the moving body information storage means 207, but in FIG. Omits the line of reference.

Next, the operation of the moving object information estimating means 259 will be described. The moving object speed estimation means 301 estimates the speed at the monitored area coordinates of each previously detected moving object from the past moving object information of the previously detected moving object stored in the moving object information storage means 207. The moving body estimated position calculating means 302 calculates the moving body information in the past from the moving body information of the moving body detected in the past stored in the moving body information storage means 207 and the speed of each moving body estimated by the moving body speed estimating means 301. The current position of each detected moving object in the coordinates of the monitored area is estimated, estimated moving object information is generated for each previously detected moving object, and the estimated current position is set and output. It should be noted that the moving body speed estimating means 301 and the moving body estimated position calculating means 302
Performs the same operation as the moving object velocity estimating means 141 and the moving body estimated position calculating means 142 in the first embodiment, and detailed description thereof will be omitted here. Therefore,
The moving body position estimating means 300 also performs the same operation as the moving body position estimating means 140 of the first embodiment, and outputs estimated moving body information in which the estimated current position of the moving body detected in the past is set.

The image coordinate calculating means 303 receives the estimated moving body information from the moving body position estimating means 300, and the moving image input means 201 of each moving body from the estimated position of each moving body calculated by the moving body position estimating means 300. The position in the image coordinate system is calculated, and the value is set as the estimated moving object information and output. The calculation is performed between the position in the surveillance area coordinates and the position in the image coordinates in the same manner as in the first embodiment (Equation 1).
4) and (Equation 15) are used.

The moving object shape estimating means 304 receives the estimated moving object information from the image coordinate calculating means 303, and the latest height of the moving object information corresponding to each number of each estimated moving object information is used as the moving object information. Read from the storage means 207,
The value is set as the estimated height of the moving body in the estimated moving body information and output.

As described above, the moving body information estimating means 2
Reference numeral 59 indicates the current position in the monitored area coordinates of each moving object and the position and height in the image coordinate system of the moving image input means 201 from the moving object information of the previously detected moving objects stored in the moving object information storage means 207. Estimate and output as estimated moving object information.

As described above, the moving object detecting means 203 in the present embodiment has the moving area detected by the moving image processing means 202 and the moving object detected in the past stored in the moving object information storage means 207. Information and moving image input means 201
Detects the moving body in the monitoring area of and calculates the moving body information,
It can be output as the output of the monitoring device in the present embodiment and can be output to the moving object information storage unit 207.

(Effect of Second Embodiment) As described above,
The distributed monitoring apparatus according to the second embodiment processes a moving image captured by a plurality of moving image input means 201, 204 by a plurality of moving image processing means 202, 205 to calculate a moving area, The moving object information is detected by the moving object information in the moving object detecting means 203 and 206, and when the moving object is detected, the moving object information of the moving object previously detected by other moving object detecting means is also detected. Even if the moving object to be monitored moves from the monitoring area of one moving image input means to the monitoring area of the other moving image input means by referring through the storage means, the estimated shape and the estimated position are used. It is possible to detect moving objects by associating moving regions with moving objects that have been detected in the past, and it is possible to perform monitoring with less error in detecting moving objects and to track complex movements of moving objects. Is possible .

Further, the moving object information storage means 207, 208
Since the moving object information of the moving object detected in the past is stored in the past and the moving object is detected by integrating the past information, even if the detection of the moving object fails temporarily, it can be supplemented. Therefore, it is possible to perform monitoring with less error in detecting the moving object.

Furthermore, each moving image input means 201, 20
4, moving image processing means 202, 205, moving object detection means 2
Outputs 03 and 206 are a moving image, a moving region, and moving object information, respectively, and since these information do not directly depend on internal processing, moving image input means, moving image processing means, moving object detection Even if the configuration or operation of any of the means is changed, the influence on the other parts can be minimized.

Also, the position of the moving object corresponding to the moving area calculated by the area position calculating means 228 of the moving image processing means 202, 205 in the monitoring area coordinates is calculated, and the position information is calculated as a plurality of moving image input means 201. , 204 in association with a moving object detected in the past from the moving image input in 204, the reliability of integration of information from a plurality of moving image input means 201, 204 and integration of past information is increased. be able to.

Further, the moving object information estimating means 259 estimates the moving object information of the moving object detected in the past, and the estimated moving object information and the moving area information output from the moving image processing means are associated with the moving object moving area. By associating with the attaching means, the reliability of the integration of information from a plurality of moving image input means and the integration of past information can be enhanced.

Also, the current position of the moving object detected in the past is estimated by the moving object position estimating means 300, and the moving object concealment determining means 281 conceals the previously detected moving object with other previously detected moving objects. By determining whether or not the moving object is detected from the estimated current position of the moving object, it is possible to prevent an error in associating the moving object with the moving area even when the corresponding moving area cannot be detected by one moving image input unit.

Further, the coordinate value in the image coordinate system of the moving image input means corresponding to the estimated current position of the moving object detected by the image coordinate calculation means 303 in the past is obtained, and the coordinate value is used by the photographing range determination means 283 in the past. Even if the moving object detected in the past goes out of the imaging range of the moving image input means, the moving object and moving area are detected by determining whether or not the moving object detected in the moving image input means is imaged. It is possible to prevent an association error with.

Also, the moving image processing evaluation means 241, 242.
By evaluating the result of the moving image processing in the moving image processing means 202, 205, and adjusting the state of the moving image input by the moving image input means 201, 204 via the moving image input adjusting means 213 by the evaluation result. Not only can the state of moving image input be maintained in a better state, but also the time and effort required for adjustment during installation can be saved.

(Modified Embodiment) In the first and second embodiments, the number of moving image input means is two, but the number is not limited to this, and the number is more than that. Also has the same effect.

Further, in the second embodiment, the number of moving object information storage means is M, but the number is not limited to this, and one moving object information storage means is shared by a plurality of moving object detection means. You may.

In the first embodiment, the image pickup ranges of both moving image input means are overlapped, and in the second embodiment, the image pickup ranges of both moving image input means are not overlapped. The present invention is not limited to this. For example, even if a part of the moving image input means is overlapped, or a case where the moving image input means is overlapped and a case where they are not overlapped are mixed, the same effect can be obtained. can get.

Further, in the first and second embodiments, the moving image input means has a television camera as a constituent element, but the invention is not limited to this, and it is an infrared camera or a night vision camera. However, the same effect can be obtained. Further, the same effect can be obtained even if a television camera, an infrared camera, a night-vision camera and the like are mixed.

The texture information is not used for the moving object information and the moving area information in the first and second embodiments, but the area texture such as the edge density of the area is used as the feature amount. If used, the moving object and the moving object can be associated with higher reliability.

Further, in the first and second embodiments, the height of the moving object is used as the shape feature amount, but the present invention is not limited to this. For example, the aspect ratio of the moving object, the moving region. Other shape features such as the circularity of
The same effect can be obtained by combining a plurality of shape feature amounts.

In the first embodiment, YUV is used as the color.
Although the color system is used, the present invention is not limited thereto, and the same effect can be obtained by using another color system, and the same effect can be obtained by combining a plurality of color systems. be able to. Furthermore, the same effect can be obtained by using a spatial distribution of colors, a histogram shape, and the like.

Further, in the first embodiment, whether or not the height of the moving area is within a preset range is used as the moving image processing evaluation value, but the present invention is not limited to this. The same effect can be obtained by using the number of numbers as the moving image processing evaluation value.

In the first embodiment, the result of comparing the number of moving bodies detected at the present time with the number detected in the past is used as the moving body treatment evaluation value, but the present invention is not limited to this. For example, the similarity between each feature amount of the moving object information and the moving area information may be used as the moving object processing evaluation value.

Further, in the first and second embodiments, the moving image processing means is provided for each moving image input means, but the present invention is not limited to this, and a moving image output from a plurality of moving image input means. The image may be processed by one moving image processing unit, and the moving region may be calculated for each moving image input unit.

Further, in the moving body motion area associating means of the moving body detecting means in the first and second embodiments, the processing by the feature associating portion is performed and then the processing by the unsupported moving object processing portion is performed. However, the present invention is not limited to this, and the moving object that is estimated not to be associated with the moving area out of the moving objects detected in the past by performing the processing of the unsupported moving object processing unit first It is also possible to adopt a configuration in which no matching is attempted. For example, even if the moving object that has been previously determined by the shooting range determination means to be outside the shooting range of the moving image input means is not associated with the moving area obtained by processing the corresponding moving image input means. Good.

[0353]

As described above, the distributed monitoring apparatus according to the present invention includes a plurality of moving image input means, a moving area corresponding to a moving object from each moving image output from each moving image input means, and its moving area. Moving image processing means for outputting moving area information consisting of feature amounts, moving body detecting means for outputting moving body information from moving area information output by the moving image processing means, and moving body output by the moving image processing means 1st structure provided with the moving body information storage means which memorize | stores information, or M moving image input means and each moving image input means, respectively, is connected to each moving image, and from each moving image which is the output, to a moving body. M moving image processing means for outputting moving area information including a corresponding moving area and its characteristic amount, and M for connecting to each moving image processing means and outputting moving object information from the output moving area information M Individual moving object detection means and M A second means comprising communication means for performing communication between the object detection means and N (where N is an integer not larger than M) moving object information storage means for storing moving object information output from the moving image processing means. With the above configuration, by integrating the result obtained by processing the moving image obtained by the plurality of moving image input means by the moving image processing means in the moving object detection means,
Even if the result of processing the moving image input by a certain moving image input means by the moving image processing means is insufficient, the moving image input by another moving image input means may be processed by the moving image processing means. This has the first effect of complementing and performing monitoring with less detection error of the moving object.

Also, by integrating the past moving body information stored in the moving body information storage means with the moving body detection means, it is possible to supplement it even if the detection of the moving body fails temporarily. It has a second effect that it is possible to perform monitoring capable of tracking the moving body even for a complicated moving body.

Furthermore, the outputs of the moving image input means, moving image processing means, and moving object detecting means are abstracted as moving images, moving regions, and moving objects, respectively, so that a plurality of types of algorithms are used by each of the above three means. And hardware can be mixed, and even if the algorithms and hardware are changed, the influence on other means can be minimized, and flexible and reliable monitoring is possible. It has the third effect that the device can be provided.

Further, in addition to the first or second configuration, the moving image processing means has a third area position calculating means for calculating the spatial position of the object corresponding to the area from the area in the image. With the configuration, by integrating the position information of each moving area calculated by the area position calculating means, the first to third effects can be obtained.

Further, in addition to the third configuration, the area position calculating means calculates plane contact point coordinate calculating means for calculating coordinates on the image corresponding to a point at which the object imaged in the image is in contact with the plane. According to the fourth configuration having the spatial coordinate calculating means for converting the coordinates on the image, which is the output of the planar contact coordinate calculating means, into the spatial coordinates of the area where the image is picked up, each of the planar contact coordinate calculating means and the spatial coordinate calculating means The position of the moving object corresponding to the moving region can be easily obtained, and the obtained first to third effects can be obtained by integrating the obtained position information.

Further, in addition to the first or second configuration, the moving object detecting means estimates moving object information of a moving object detected in the past, and moving object information estimating means,
A moving object that estimates which moving area the moving object detected in the past corresponds to, from the current moving object information estimated by the moving object information estimating means and the moving area information output from the moving image processing means. According to the fifth configuration including the moving area associating unit, the moving object information estimating unit estimates the moving object information, and the moving object moving area associating unit estimates the estimated moving object information and the current moving image processing. By correlating with the moving area information which is the output of the means, the moving object can be tracked and the first to third effects can be obtained.

Furthermore, in addition to the third or fourth configuration, the moving object detecting means estimates moving object information of a moving object detected in the past, and moving object information estimating means,
A moving object that estimates which moving area the moving object detected in the past corresponds to, from the current moving object information estimated by the moving object information estimating means and the moving area information output from the moving image processing means. According to the sixth configuration having the moving area associating unit, the moving object information estimating unit estimates the moving object information, and the moving object associating unit estimates the moving object information and the current moving image processing unit. By associating the moving body information with the moving area information that is the output of, the moving body is associated with the moving area, and the current position of the moving body is estimated from the result using the position information of the corresponding moving area to track the moving body. By doing so, the first to third effects can be obtained.

Further, in addition to the sixth structure, the seventh structure has a moving body position estimating means for estimating the current position of the moving body detected in the past by the moving body information estimating means. By correlating the moving object and the moving object detected in the past by the moving object moving object matching unit using the estimated current position of the moving object detected in the past, the first to third effects can be obtained. Obtainable.

Further, in addition to any one of the fifth to seventh configurations, the moving image processing means has area color calculation means for calculating the color of the object corresponding to the area from the area in the image,
With the eighth configuration having the animal body color estimating means for estimating the color of the moving body detected in the past by the moving body information estimating means, the area color calculating means calculates the color of the moving area, and the moving body color estimating means further calculates the color. More reliable by estimating the color of the moving object and using the similarity between the moving area color and the moving object color to associate the estimated moving object information with the moving area information in the moving object moving area associating means. The first to third effects can be obtained by associating a moving object with a highly movable body and a moving area.

Further, in addition to any one of the fifth to seventh configurations, the moving image processing means has area shape calculation means for calculating the shape of the object corresponding to the area from the area in the image, With the ninth configuration including the moving object shape estimating unit that estimates the moving object shape previously detected by the moving region associating unit, the region shape calculating unit calculates the shape of the moving region,
Further, the shape of the moving body is estimated by the moving body shape estimating means, and the degree of similarity between the moving area shape and the moving body shape is associated with the estimated moving body information and moving area information in the moving body moving area associating means. By using, it is possible to associate the moving body with the moving body with higher reliability, and to obtain the first to third effects.

Furthermore, in addition to any of the fifth to seventh configurations, the moving image processing means has area texture calculation means for calculating the texture of the object corresponding to the area from the area in the image, According to the tenth configuration having the moving object texture estimating means for estimating the texture of the moving object detected in the past by the information estimating means, the texture of the moving area is calculated by the area texture calculating means, and the moving object texture estimating means further calculates the moving body texture. Of the texture of the moving body and the texture of the moving body to correlate the moving body information and the moving body information estimated by the moving body moving area associating unit By associating a high moving object with a moving area, the first to the third
The effect of can be obtained.

Further, in addition to any of the seventh to tenth configurations, the object moving area associating means stores the position of the object to be hidden in the area imaged by the moving image inputting means. Storage means, moving image input means position storage means for storing the position of the moving image input means, and the current position of the moving body estimated by the moving body position estimation means is concealed by the concealed object stored in the concealed object position storage means According to the eleventh configuration including the concealment determination unit that determines whether or not the concealment object stored in the concealment object position storage unit conceals the estimated current position of the moving object. It is determined by the concealment determining means using the position of the moving image inputting means stored in the moving image inputting means position storing means, and when the moving object is hidden by the concealing object, the corresponding moving area is determined. Exists By not performing the association between the moving region the Most determines prevent association errors between the moving region of the animal body, it is possible to obtain the third effect from the first.

Further, in addition to any one of the seventh to tenth constitutions, the moving body motion area associating means calculates the moving body from the current position of the moving body detected in the past estimated by the moving body position estimating means. According to the twelfth configuration having a moving object hiding determination unit that determines whether or not is hidden by another moving object detected in the past, the estimated current position of the moving object detected in the past estimated by the moving object position estimating unit. From the above, the moving object hiding determination means determines whether or not the moving object is hidden by another previously detected moving object, and if the moving object is hidden by another moving object, the corresponding moving area exists. By not determining that the moving area is not associated with the moving area, an error in associating with the moving area of the moving object can be prevented, and the first to third
The effect of can be obtained.

Further, in addition to any one of the seventh to tenth constitutions, there is further provided a first image coordinate calculating means for calculating to which pixel of a moving image of each moving image processing means a point in a certain space corresponds. With the configuration of 13, the image coordinate calculation means obtains the coordinate position corresponding to the estimated current position of the previously detected moving body estimated by the moving body position estimating means in each moving image of each moving image input means, and the moving body movement is calculated. It is determined whether or not the moving object detected in each of the past is captured in each of the moving images by the area association unit, and it is not determined that the moving object detected in the past is captured in each of the moving images. If it is determined that the corresponding moving area does not exist and the moving area is not associated with the moving area, an error in associating with the moving area of the moving object is prevented, and the first to third effects are obtained. You can

Further, in addition to any of the seventh to tenth constitutions, the moving object input means position storing means for storing the position of each moving image input means, and each moving image With the fourteenth configuration having distance threshold storage means for storing the distance threshold corresponding to the input means, the estimated current position of the moving object detected in the past estimated by the moving body position estimating means and the moving image input means position storing means. The distance to each position of the moving image processing means stored in is calculated, and if this distance is larger than the distance threshold stored in the distance threshold storage means, the moving object is outside the range detected as the moving area. By determining that there is no correspondence between the moving object and the moving object detected in the past by the moving object moving object matching means,
It is possible to prevent an error in associating with a moving region of a moving object and obtain the first to third effects.

Further, in addition to any one of the first to fourteenth constructions, the moving image input adjusting means for changing the moving image input state of the moving image input means, and the calculation result of the moving area information of the moving image processing means. With the fifteenth configuration having the moving image processing evaluation means for calculating the evaluation value of, the result of processing each moving image is evaluated by the moving image processing evaluation means, and the state of moving image input is changed as a result of this evaluation. If you decide it is better,
By changing the state of the moving image input by the moving image input adjusting means, the state of the moving image input can be maintained in a better state. In addition to the first to third effects, labor saving of the adjustment at the time of installation can be achieved. At the same time, it is possible to obtain the fourth effect that it is possible to perform monitoring with a high degree of flexibility that follows changes in the environment and with a small detection error of the moving object.

Further, in addition to any one of the first to fourteenth configurations, the moving image input adjusting means for changing the moving image input state of the moving image input means, and the moving object information output from the moving object detecting means. A sixteenth configuration having a moving object detection and evaluation means for calculating an evaluation value of, and evaluating the result of detecting a moving object by the moving object detection and evaluation means, and changing the state of moving image input as a result of this evaluation. If it is determined that the moving image input adjusting means changes the moving image input state, the moving image input state can be maintained in a better state, and the first to fourth effects are obtained. be able to.

Further, in addition to any one of the first to fourteenth constructions, the moving image processing adjusting means for changing the calculation state of the moving area information of the moving image processing means, and the animal output from the moving object detecting means. With the seventeenth configuration having a moving body detection and evaluation means for calculating an evaluation value of the body detection result, the moving body detection and evaluation means evaluates the detection result of the moving body, and the result of this evaluation, the state of moving image processing. If it is determined that the moving image processing state should be changed, the moving image processing state can be maintained in a better state by changing the moving image processing state by the moving image processing adjustment means. The effect of can be obtained.

Further, in addition to the seventeenth structure, the eighteenth structure having a moving image input adjusting means for changing the moving image input state of the moving image input means makes it possible to detect the moving body by the moving body detection evaluating means. Evaluate the result, the result of this evaluation,
When it is determined that the moving image processing state should be changed, when it is determined that the moving image processing adjusting unit changes the moving image processing state and the moving image input state is changed. By changing the state of moving image input by the moving image input adjusting means, the state of moving image input and moving image processing can be maintained in a better state, and the first to fourth effects can be obtained. .

Further, in addition to the first to fourteenth constructions, the moving image processing adjusting means for changing the calculation state of the moving area information of the moving image processing means and the moving image input state of the moving image input means are changed. According to the nineteenth configuration having the moving image input adjusting means, the moving image processing evaluation means evaluates the processing result of each moving image, and as a result of this evaluation, it is determined that the state of the moving image processing should be changed. In this case, by changing the moving image processing state by the moving image processing adjusting means, the moving image processing state can be maintained in a better state, and the first to fourth effects can be obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a distributed monitoring apparatus according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a moving image input means 1 and a moving image processing means 2 in the first embodiment.

FIG. 3 is a block diagram showing a configuration of a moving image processing means 2 in the first embodiment.

FIG. 4 is a block diagram showing a configuration of a moving body detection unit 3, a moving body information storage unit 4, and a moving body detection evaluation unit 5 in the first embodiment.

FIG. 5 is a block diagram showing a configuration of a region shaping unit 26 in the first embodiment.

FIG. 6 is a block diagram showing a configuration of a region position calculating means 28 in the first embodiment.

FIG. 7 is a block diagram showing a configuration of a moving area information calculation unit 29 in the first embodiment together with related circuits.

FIG. 8 is a block diagram showing a configuration of a feature association unit 53 in the first embodiment.

FIG. 9 is an unsupported moving object processing unit 54 in the first embodiment.
Block diagram showing the configuration of

FIG. 10 is a moving object information calculation unit 58 in the first embodiment.
Block diagram showing the configuration of

FIG. 11 is a moving object information estimating means 5 in the first embodiment.
Block diagram showing the configuration of No. 7

FIG. 12 is a block diagram showing the configuration of a distributed monitoring apparatus according to a second embodiment of the present invention.

FIG. 13 is a moving image input means 201 in the second embodiment.
Block diagram showing the configuration of

FIG. 14 is a moving image processing means 202 in the second embodiment.
Block diagram showing the configuration of FIG.

FIG. 15 is a view showing a moving object detecting means 203 in the second embodiment.
Block diagram showing the configuration of the

FIG. 16 is a motion area information calculation unit 22 according to the second embodiment.
9 is a block diagram showing the configuration of FIG. 9 together with related circuits.

FIG. 17 is a feature association unit 252 according to the second embodiment.
Block diagram showing the configuration of

FIG. 18 is an unsupported moving object processing unit 2 in the second embodiment.
Block diagram showing the configuration of 54

FIG. 19 is a moving object information calculation unit 25 according to the second embodiment.
Block diagram showing the configuration of No. 7

FIG. 20 is a view showing a moving object information estimating means 2 in the second embodiment.
Block diagram showing the configuration of 59

[Explanation of symbols]

1 Moving image input means 2 Moving image processing means 3 Moving object detection means 4 Moving object information storage means 5 Moving object detection and evaluation means 6 LAN 7 Moving image input means 8 Moving image processing means 11 TV camera 12 A / D converter 13 Moving image input adjusting means 14 Aperture adjustment device 15 Aperture adjustment amount calculation circuit 21 Communication means 22 frame buffer 23 frame buffer 24 Difference calculation circuit 25 Threshold processing circuit 26 Area shaping section 27 Labeling processing circuit 28 area position calculation means 29 Motion Area Information Calculation Unit 30 Moving image processing adjusting means 31 Threshold calculator 32 Object detection evaluation value decoding circuit 33 Communication means 34 Communication means 35 Communication means 51 means for associating animal body movement areas 52 motion area information decoding circuit 53 Feature Matching Unit 54 Unsupported moving object processing unit 55 Unsupported motion area processing unit 56 Correlation Confirmation Unit 57 Animal body information estimation means 58 Animal Body Information Calculation Department 61 Moving object detection evaluation circuit 62 Object detection evaluation value coding circuit 71 Moving Area Information Encoding Circuit 72 Moving image processing evaluation means 81 Expansion processing circuit 82 Thinning circuit 83 Noise removal circuit 91 circumscribed rectangle calculation circuit 92 Floor coordinate detection circuit 93 space coordinate calculation means 94 Plane contact point coordinate calculation means 101 circumscribed rectangle calculation circuit 102 area color calculation means 103 area shape calculation means 104 Moving Area Information Generation Unit 111 Distance correspondence circuit 112 shape matching circuit 113 Chromaticity correspondence circuit 114 association determination circuit 121 Concealed object position storage means 122 Concealment judging means 123 Moving image input means Position storage means 124 Object Concealment Judgment Means 125 shooting range determination means 126 distance threshold storage means 127 Moving area detection range determination means 131 Animal body number determination unit 132 Moving object position determination unit 133 Moving object shape determining unit 135 Animal body color determination unit 136 Animal body information determination unit 140 Moving object position estimating means 141 Animal Body Speed Estimating Means 142 Moving Object Estimated Position Calculation Means 143 image coordinate calculation means 144 Moving object shape estimating means 145 Animal Body Color Estimating Means 201 Moving image input means 202 Moving image processing means 203 Object detection means 204 Moving image input means 205 Moving image processing means 206 Object detection means 207 Moving object information storage means 208 Animal Body Information Storage Means 209 LAN 211 TV camera 212 A / D converter 213 Moving image input adjusting means 214 Aperture adjustment device 215 Aperture adjustment amount calculation circuit 222 frame buffer 223 frame buffer 224 Difference calculation circuit 225 threshold processing circuit 226 Area shaping section 227 Labeling processing circuit 228 area position calculation means 229 Motion Area Information Calculation Unit 241 Moving image processing evaluation means 242 Moving image processing evaluation means 251 Animal Body Motion Area Correlating Means 252 feature association unit 254 Unsupported moving object processing unit 255 Unsupported motion area processing unit 256 association determination unit 257 Animal Body Information Calculation Department 258 Moving object information encoding circuit 259 Animal body information estimation means 260 Moving Object Information Decoding Circuit 261 circumscribed rectangle calculation circuit 262 area shape calculation means 271 Distance correspondence circuit 272 Shape correspondence circuit 273 Correlation determining circuit 281 Moving object hiding judgment means 282 Moving image input means Position storage means 283 Shooting range determination means 291 Animal body number determination unit 292 Moving object position determination unit 293 Moving object shape determining unit 300 Moving object position estimation means 301 Animal speed estimation means 302 Estimated position of moving body 303 image coordinate calculation means 304 Moving object shape estimating means

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Toshi Fuji Oka 3-10-1 Higashisanda, Tama-ku, Kawasaki City, Kanagawa Prefecture Matsushita Giken Co., Ltd. 3-10-10 Mita Matsushita Giken Co., Ltd. (72) Inventor Masato Mori 4-1 Egasakicho, Tsurumi-ku, Yokohama-shi, Kanagawa Tokyo Electric Power Co., Inc. System Research Laboratory (56) Reference JP-A-1-140391 (JP, A) JP-A-2-82375 (JP, A) JP-A-4-150592 (JP, A) JP-A-8-16790 (JP, A) JP-A-7-262462 (JP, A) Kaihei 4-261289 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G08B 21/00 H04N 7/18

Claims (25)

(57) [Claims] 1. A plurality of moving image inputting means, and moving image processing means for outputting moving area information including a moving area corresponding to a moving object and its feature amount from each moving image output from each of the moving image inputting means. If, comprising: a moving object detecting means for outputting a moving object information from the moving region information which is the output of the moving picture processing unit, and a moving object information storing means for storing the moving object information, which is the output of the moving object detecting means , The moving object detecting means integrates the moving area information in the moving images obtained by the plurality of moving image inputting means, and also integrates the past moving body detection results stored in the moving body information storing means. A distributed monitoring device for detecting a moving object. 2. An image pickup range at least partially overlaps with M moving image input means, M moving image processing means, and each of the moving image input means has one moving image input means. 2. The distributed monitoring apparatus according to claim 1, wherein the moving image processing means is connected. 3. M moving image input means and moving area information consisting of a moving area corresponding to a moving object and its feature amount from each moving image output from each of the moving image inputting means. M moving image processing means for outputting, M moving body detecting means connected to each of the moving image processing means and outputting moving body information from moving area information which is the output, and each moving body detecting means Communication means for communicating between the means,
A moving image provided with N (where N is an integer not larger than M) moving object information storing means for storing moving object information output from each of the moving object detecting means, and to which the certain moving object detecting means is connected. The moving body information which is the output of the processing means and the moving body information which is the output of the other moving body detecting means are integrated, and the past moving body detection result stored in the moving body information storage means is also integrated to obtain the moving body information. A distributed monitoring device characterized by detecting the following. 4. The method of claim 1 or請 imaging range of the moving image input means, characterized in that the overlap at least a portion
The distributed monitoring device according to claim 3 . 5. The distributed monitoring apparatus according to claim 1, wherein the moving image input means includes a television camera. 6. The distributed monitoring apparatus according to claim 1, wherein the moving image input means includes an infrared camera. 7. The distributed monitoring apparatus according to claim 1, wherein the moving image input means includes a night vision camera. 8. The moving image processing means has area position calculating means for calculating the spatial position of an object corresponding to the area from the area in the image, and the moving area information indicates the moving object of the moving object corresponding to each moving area. The distributed monitoring device according to any one of claims 1 to 7, wherein the distributed monitoring device includes each moving body position. 9. An area position calculating means calculates plane contact point coordinate calculating means for calculating coordinates on an image corresponding to a point at which an object imaged in the image is in contact with a plane, and an output of the plane contact point coordinate calculating means. 9. The distributed monitoring apparatus according to claim 8, further comprising: spatial coordinate calculation means for converting the coordinates on the image into the spatial coordinates of the area where the image is captured. 10. An animal body information estimating means for estimating a current animal body information of an animal body detected in the past, and a current animal body information and a moving image estimated by the animal body information estimating means. Moving object information which is output from the image processing means and has a moving object associating means for estimating which moving object the moving object detected in the past corresponds to, and the moving object detecting means, The distributed type according to any one of claims 1 to 7, wherein the distributed type has a function of updating the moving object information of the moving object detected in the past from the result of the matching by the moving body motion area associating means. Monitoring equipment. 11. An animal body information estimating means for estimating current animal body information of an animal body detected in the past, and a current animal body information and a moving image estimated by said animal body information estimating means. Moving object information which is output from the image processing means and has a moving object associating means for estimating which moving object the moving object detected in the past corresponds to, and the moving object detecting means, dispersing the results associated with the moving object moving region correlating means in the past and having the function of updating the moving object information including the current position of the detected moving object according to claim 8 or claim 9 Symbol mounting Type monitoring device. 12. The moving body information estimating means has moving body position estimating means for estimating the current position of a moving body detected in the past, and the moving body motion area associating means estimates by the moving body position estimating means. Using the distance between the current position of the moving object and the position of the moving object in the moving area information output from the moving image processing means, it is determined which moving area the moving object detected in the past corresponds to. distributed monitoring system of claim 1 1 Symbol mounting and having a function of estimating. 13. The moving body position estimating means outputs the estimated position when the moving body moves from the past fixed position at the speed of the past moving body.
The distributed monitoring device described. 14. The moving image processing means has area color calculation means for calculating the color of the object corresponding to the area from the area in the image, and the moving object information estimating means has the color of the moving object detected in the past. A moving object information which is an output of the moving image processing means and the current color of the moving object estimated by the moving body color estimating means. 14. A moving object detected in the past is estimated to correspond to a current moving region using a degree of similarity between the moving object and the color of the moving object in FIG. The distributed monitoring device according to 1. 15. The moving image processing means has area shape calculating means for calculating the shape of an object corresponding to the area from the area in the image, and the moving object information estimating means has the shape of the moving object detected in the past. A moving object information which is an output of the moving image processing means and the current shape of the moving object estimated by the moving object shape estimating means. 14. Which of the present moving regions the moving object detected in the past corresponds to is estimated using the similarity between the moving object and the shape of the moving object in FIG. The distributed monitoring device according to 1. 16. The moving image processing means has area texture calculation means for calculating the texture of an object corresponding to the area from the area in the image, and the moving object information estimating means has texture of the moving object detected in the past. And a moving body area estimation unit that estimates the moving body moving area associating unit, and moving body area information that is an output of the current texture of the moving body estimated by the moving body texture estimating unit and the moving image processing unit. 14. A moving object detected in the past is estimated to correspond to a current moving region by using a similarity between the moving object and the texture of the moving object in FIG. The distributed monitoring device according to 1. 17. A moving object moving area associating means stores a concealing object position storing means for storing a position of a concealing object in an area imaged by the moving image inputting means, and a position of the moving image inputting means. Whether or not the current position of the moving object estimated by the moving image input means position storage means and the moving object position estimation means is hidden from the moving image input means by the hidden object stored in the hidden object position memory means. It has a concealment judging means for judging, and when it judges that the moving object detected in the past corresponds to the current moving area and the moving object detected in the past is concealed by the concealing object, it corresponds. The distributed monitoring device according to any one of claims 12 to 16, wherein the distributed monitoring device is not attached. 18. A moving object detected by the moving image inputting means in the past from the current position of the moving object detected in the past estimated by the moving body position estimating means by the moving object moving area associating means. At the time of estimating whether the moving object corresponding to the moving object detection unit in the past, which has a moving object hiding determination unit that determines whether or not the moving object is previously hidden, The distributed type according to any one of claims 12 to 16, wherein when the moving object detected in the past is concealed by another moving object detected in the past, no association is performed. Monitoring equipment. 19. An image coordinate calculating means for calculating which pixel of a moving image of each moving image processing means corresponds to a point in a certain space, and the moving body moving area associating means has detected in the past. If it is determined that the current position of the moving object, which is the output of the moving object position estimating means, is outside the shooting range of the moving image input means when estimating which moving area the moving object corresponds to, The distributed monitoring apparatus according to any one of claims 12 to 16, wherein the distributed monitoring apparatus is not associated with the moving area of the moving image input means. 20. The moving object input means position storage means for storing the position of each moving image input means, and the distance threshold storage means for storing the distance threshold corresponding to each moving image input means. The present position of the moving object detected in the past estimated by the moving object position estimating unit and the moving image input unit when estimating which moving region the moving object detected in the past corresponds to If the distance to the position of a moving image input means stored in the position storage means exceeds the corresponding distance threshold of the distance threshold storage means, the moving image input means should not be associated with the moving area. The distributed monitoring device according to any one of claims 12 to 16. 21. A moving image input adjusting means for changing a moving image input state of the moving image input means, and a moving image processing evaluation means for calculating an evaluation value of a calculation result of moving area information of the moving image processing means. Then, the state of the moving image input is changed by the moving image input adjusting means by using the evaluation value which is the output of the moving image processing evaluation means.
The distributed monitoring device according to any one of 0. 22. A moving image input adjusting means for changing a moving image input state of the moving image input means, and a moving body detection evaluating means for calculating an evaluation value of moving body information output from the moving body detecting means. Then, the state of the moving image input is changed by the moving image input adjusting means by using the evaluation value which is the output of the moving object detecting and evaluating means.
The distributed monitoring device according to any one of 0. 23. A moving image processing adjusting means for changing a state of calculation of moving area information of the moving image processing means, and a moving object detection evaluation for calculating an evaluation value of a detection result of the moving object output from the moving object detecting means. 21. means for changing the calculation state of the moving area information by the moving image processing adjusting means using an evaluation value output from the moving object detection and evaluation means. The distributed monitoring device according to any one of 1. 24. A moving image input adjusting means for changing a moving image input state of the moving image input means, wherein the moving image input adjusting means uses the evaluation value output from the moving object detection evaluating means. 24. The distributed monitoring device according to claim 23, wherein the input state is changed. 25. A moving image processing adjusting means for changing the calculation state of the moving area information of the moving image processing means, and a moving image processing evaluation means for calculating an evaluation value of the calculation result of the moving area information of the moving image processing means. 21. The method according to claim 1, further comprising: and changing the calculation state of the moving area information by the moving image processing adjusting means using the evaluation value output from the moving image processing evaluating means. The distributed monitoring device according to any one of claims.