JPH06301782A - Monitor device - Google Patents

Abstract

PURPOSE:To improve the detection precision of the monitor device which detects and tracks a moving body entering an object area to be monitored by processing the output image of an image pickup device 4. CONSTITUTION:History information showing the track of the moving body is generated by a moving body detection part 14 and a moving body tracking part 15. A variable gazed area setting part 13 finds past curvature variation and variation in moving speed on the basis of the history information and estimates the size of the moving body in an image at next time. Then the size of a gazed area as an image processing area is varied according to the estimated size. This processing is repeated at specific intervals of time.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to monitoring devices, and in particular to
The present invention relates to an apparatus that detects and tracks a moving object to be monitored by performing image processing on a monitoring target area captured by an image capturing apparatus.

[0002]

2. Description of the Related Art In recent years, as a safety measure, computer image processing technology has been used to automatically monitor illegal intruders in places where entry is prohibited and illegal parking in prohibited parking areas. There is a growing interest in the monitoring equipment that was used. This type of monitoring apparatus generally performs image processing such as image conversion processing and feature extraction processing on a two-dimensional image input from an image pickup apparatus such as a television camera, and a person or vehicle invading the monitoring target area. Is detecting a moving object. In addition, by associating the position of the moving object at the previous time with the position at the current time, the trajectory of the object is detected and its motion is analyzed. Furthermore, paying attention to the existence area of the moving object, the size, shape, and
By extracting features unique to an object such as color, it is possible to determine what the moving object is.

Regarding such a conventional technique, for example, refer to the description of "Motion Detection of Multiple Objects Using Multiprocessor System" (Report of IPSJ Computer Vision Research Group, March 1992). You can

FIG. 6 is a schematic diagram showing a concrete example of the configuration of a conventional monitoring device of this type. A computer 1, a display unit (CRT) 2, a memory 3, and an image pickup device 4 such as a television camera are shown.
Composed of and. The computer 1 includes an arithmetic processing unit (CPU) 11, an image input unit 12, and a gazing area setting unit 1.
6, a moving object detection unit 14, and a moving object tracking unit 15 are provided, and these units are connected by a data bus 17.

In the monitoring device having the above structure, the color grayscale image of the monitoring target area input from the image pickup device 4 is
It is converted into a digital frame image by the image input unit 12,
It is input to the moving object detection unit 14. Moving object detector 14
Then, the moving objects in the images input one after another are detected, and the position on the image is calculated.

A technique for detecting such a moving object is described, for example, in "Technology for Extracting Moving Object Image" by Kawabata, Tanifuji, and Morooka (Information Processing Society of Japan, April 1987 issue). As described above, a background difference processing technique for performing a difference process regarding the brightness value of the input image and the background image in which the moving object does not exist is applied.

The position of the moving object calculated in this manner is accumulated in the memory 3, displayed on the display unit 2, and further guided to the moving object tracking unit 15.

The moving object tracking section 15 associates the moving objects with reference to the past trajectory of the moving objects. In this association, specifically, the object having the smallest detected position difference between the moving objects between successive times is regarded as the same object. The moving object tracking unit 15 also predicts the detected position of the moving object at the next time based on the past trajectory.

As a technique for predicting the detection position, for example, "The Application of a Hybrid Tracking Algorithm to Fletcher, Warwick, Mitchell"
Motion Analysis ”(Proc.of CVPR'91, 1991)
A position estimation technique using smoothness of motion, etc., as described in 1) is applied.

The predicted position of the moving object is guided to the gaze area setting unit 16, and the area where the moving object is likely to exist at the next time, that is, the gaze area for performing the image processing at the next time is determined. Is set. 7 (a) (b) (c)
3A and 3B are diagrams illustrating an example of setting the gaze area when the time t is k, k + 1, and k + 2, respectively, in which 30a to 30c denote moving objects to be monitored, and 34 denotes a gaze area. This gaze area 34
Is generally determined in advance in consideration of the dimensions of the moving objects 30a to 30c on the image, and the moving object detection process at the next time is performed within the gaze area 34.

By repeating the above processing, the trajectory of the moving object in the image is detected without interruption, and the tracking of the moving object becomes possible.

[0012]

As described above, in the conventional monitoring device, the size of the gaze area 34 is fixed to a predetermined size, as shown in FIGS. 7 (a) to 7 (c). As described above, even if the size of the moving objects 30a to 30c changes on the image due to the distance or deformation of the moving objects
The moving object detection process is always performed on the gaze area 34 having the same size. Considering such dimensional changes on the images of the moving objects 30a to 30c, the size of the gaze region 34 is
It is generally set to a rather large value. For this reason, for example, when the moving object is distant and the size of the moving object is observed to be small, a gazing area 34 larger than the area where the existence of the moving object is predicted in the real world is set. As shown in (c), there is a high possibility that the object 30d other than the moving object to be monitored is included in the gazing area 34. On the contrary, when the size of the gaze area 34 is set to be small, when the moving object makes a complicated motion and the accuracy of the position prediction deteriorates, there is a high possibility that the gaze area 34 is not included in the gaze area 34. That is, in either case, there is a problem that the detection processing of the moving object to be monitored cannot be performed accurately.

The present invention has been made in view of the above problems, and an object thereof is to achieve accurate measurement even if the size of a moving object on the image changes or the accuracy of position prediction deteriorates. It is to provide a monitoring device capable of performing detection processing.

[0014]

The structure of the present invention for achieving the above object is an apparatus for monitoring a specific moving object invading into a monitoring target area, and an image pickup apparatus for picking up an image of the monitoring target area. A moving object detecting means for detecting the moving object by processing the captured image at predetermined time intervals, and a gaze area for setting a gaze area for the image processing at a detection predicted position of the moving object at the next time. In a monitoring device comprising a setting means, first means for generating history information representing the trajectory of the detected moving object based on the position and dimensions on the continuous past and present time images of the moving object, A second evaluation value is calculated from this history information, which indicates the complexity of the past trajectory, and a second time dimension of the moving object is estimated based on the evaluation value and the current time dimension of the moving object. And stage, depending on the estimated size is characterized by having a third means for updating the size of the fixation region in the next time, the.

Further, another configuration of the present invention for achieving the above object is, in the above monitoring apparatus, further comprising: tracking the moving object by associating the history information and detection positions between successive current times. It is characterized by having four means.

The evaluation value is, for example, a value obtained by combining the curvature change of the past trajectory of the moving object and the change of the moving speed by multiplying each by a predetermined coefficient. Further, the characteristics of the evaluation function can be freely set according to the type of the moving object.

[0017]

In the monitoring apparatus of the present invention, the image pickup apparatus is installed so that the monitoring target area is included in the field of view. Then, a moving object that enters the monitoring target area is detected from the gaze area on the captured image. By continuously detecting the position of the moving object and associating it with the past position, the locus of movement of the object can be known. The first means generates history information representing this locus and sends it to the second means. The second means calculates an evaluation value representing the complexity of the past trajectory of the moving object from the history information. This evaluation value is, for example, a value obtained by combining the change in the curvature of the past trajectory of the moving object and the change in the moving speed by multiplying each by a predetermined coefficient, and the movement of the object up to the current time can be digitized. it can. By reflecting the evaluation function based on this evaluation value on the current size of the moving object, the size of the moving object at the next time can be estimated. The third means changes the size of the gaze area on the image according to the estimated size. As a result, even when a moving object moves, deforms, or makes a complicated movement in the real world, the size of the gazing area automatically changes in response to these factors, and accurate detection of the moving object to be monitored is possible. Processing becomes possible.

Further, the moving object can be tracked by associating the history information of the moving object with the detected position of the current time by the fourth means.

[0019]

Embodiments of the present invention will now be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing the structure of a monitoring apparatus according to an embodiment of the present invention. The same functions as those of the conventional apparatus shown in FIG. 6 are designated by the same reference numerals.

Referring to FIG. 1, the monitoring apparatus according to this embodiment includes a computer 1, a display unit (CRT) 2 for displaying information such as characters and images, a memory 3 for storing information such as characters and images, and a television. An imaging device 4 such as a camera is provided. The computer 1 includes a CPU 11 and an image input unit 1.
2, a variable gaze area setting unit 13, a moving object detection unit 14, and a moving object tracking unit 15 are provided, and these units are connected by a data bus 17.

The display unit 2 is composed of, for example, a CRT display device, and the memory 3 is composed of a main storage device of a computer and an auxiliary storage device such as a magnetic disk device or an optical disk device. The imaging device 4 is arranged such that its field of view faces the monitoring target area, as in the conventional device. The image input unit 12 converts the color grayscale image input from the image pickup device 4 into a digital frame image.

The variable gazing area setting unit 13 changes the size of the gazing area where the moving object will exist at the next time at a predetermined time interval according to the size change on the image and the complexity of the movement. It is an update. This will be described later. The moving object detection unit 14 detects a moving object in the gaze area (moving object detection means) as in the conventional device.

The moving object tracking unit 15 generates history information representing the locus of the detected moving object, and regards the object having the smallest distance difference between the position at the previous time and the position at the current time as the same object. , Tracking processing of moving objects. Further, similarly to the conventional device, the position of the moving object on the image at the next time is predicted.

Next, the operation of the monitoring device configured as described above will be specifically described with reference to FIGS. Figure 2
Is a flow chart showing a processing procedure of the present embodiment.
Represents the step of each process.

In the monitoring apparatus of this embodiment, the images in the visual field picked up by the image pickup apparatus 4 are sequentially fetched by the computer 1 and sent to the data bus 17 via the image input section 12. The moving object detection unit 14 detects the existence area of the moving object to be monitored from the gaze area on the image, and as shown in FIG. 3, the barycentric position (TRx
The dimensions (Bx (tk), By (tk)) of the (tk), TRy (tk)) 31 and the circumscribed rectangle 32 are calculated. This center of gravity position 3
The dimensions of 1 and the circumscribed rectangle 32 are stored in the memory 3 (S21). The method of detecting a moving object is as described above.

The moving object tracking unit 15 obtains the continuous past and present time (tk) barycentric positions of the moving object by accessing the memory 3, and generates history information representing the detected trajectory of the moving object. Do (first method).
This history information is stored in the memory 3. The moving object tracking unit 15 also regards a moving object with the smallest detected position difference between successive times as the same object, and performs the tracking process of the object (fourth means). Further, the center of gravity position (TRx (tk + 1), TRy (tk + 1)) of the moving object at the next time (tk + 1) is predicted based on the history information (S22). The method of predicting the position of the center of gravity is also as described above.

The variable gaze area setting unit 13 acquires the history information of the moving object and the size of the circumscribed rectangle 32 at the current time by accessing the memory 3 (S23).
Next, in order to obtain an evaluation value (COT) that indicates the degree of complexity of the past trajectory that affects the accuracy of position prediction, a change in the curvature of the past trajectory (dC / dt) and a change in the velocity are acquired from the history information. Calculate (dV / dt) (second means: S
24). This is because in a general position prediction method, the accuracy of position prediction is highest when the moving object is performing a uniform linear motion, and the position prediction becomes difficult when the object curves or changes its speed. It is decided in consideration of the characteristics. This evaluation value (COT) is, for example, the following expression 1
It is obtained by the formula.

[0029]

COT = Kc (dC / dt) + Kv (dV / dt) where Kc and Kv are heuristically determined parameters depending on the type of the moving object to be monitored, and 0 to 1 respectively. Takes a value between.

Evaluation value (COT) calculated as described above
And a circumscribed rectangle 3 on the image of the moving object at the current time (tk)
2 is used to estimate the size of the region where the moving object exists at the next time (tk + 1) (S25). Specifically, this is calculated by the following formula 2 (PBx
It is realized by calculating (tk + 1), PBy (tk + 1)).

[0031]

[Formula 2] PBx (tk + 1) = Bx (tk) · ΔBx · (1 + F
(COT)) PBy (tk + 1) = By (tk) · ΔBy · (1 + F (CO
T)) However, ΔBx = Bx (tk) / Bx (tk−1) ΔBy = By (tk) / By (tk−1) where the function F (COT) is uniquely determined by the evaluation value (COT). Is an evaluation function that is output, and its output value takes a value from 0 to 1. FIG. 4 shows an example of the characteristic. Note that this characteristic can be freely set according to the type of moving object to be monitored. That is, the evaluation value (CO
The relationship of the output value with respect to T) can be arbitrarily set according to the motion of a human being, a vehicle, or other moving objects.

Finally, the next time (t
The circumscribed rectangle having the size calculated in S25 is arranged at the position of the center of gravity of the moving object in (k + 1) (third means). As a result, a new gaze area (hereinafter referred to as a variable gaze area) at the next time is set (S26).

FIGS. 5A, 5B, and 5C are diagrams showing examples of setting the variable gaze area when the time t is k, k + 1, and k + 2, respectively. 7A and 7B shown in FIG.
It corresponds to (c). In the figure, 33a to 33c indicate variable gaze areas updated every time. As is clear from this figure, the positions and dimensions of the variable gaze areas 33a to 33c are automatically updated at predetermined time intervals by repeating the above processing. Therefore, as shown,
Even if the moving objects 30a to 30c move or deform in the real world to change the size on the image, or if the accuracy of the position prediction deteriorates due to complicated movement, it is variable according to these factors. Since the sizes of the gaze areas 33a to 33c are sequentially changed, it is possible to effectively prevent erroneous detection due to omission of detection of the moving objects 30a to 30c and presence of the object 30d that is not a monitoring target.

Although the present invention has been specifically described with reference to the above embodiment, it is needless to say that the present invention is not limited to this embodiment and can be variously modified without departing from the scope of the invention.

[0035]

As described in detail above, the monitoring apparatus of the present invention generates history information representing the trajectory of a moving object to be monitored, and calculates an evaluation value representing the complexity of the past trajectory from this history information. At the same time, the size of the moving object at the next time is estimated based on this evaluation value and the detected size at the current time, and the size of the gazing area is updated according to the estimated size. Even if the size on the image changes due to movement or deformation in the world, or if the accuracy of position prediction deteriorates due to complicated movement, the gaze area that performs image processing in response to these factors The size is updated automatically. Therefore, there is an effect that omission or erroneous detection of a moving object is prevented.

Further, as described above, there is an effect that the tracking of the moving object becomes accurate by associating with the detection position between consecutive times of the moving object detected in the gazing area which is updated every moment. The monitoring accuracy is significantly improved compared to the conventional device.

The present invention can be used in a wide range such as detection of illegal intruders, detection of illegally parked vehicles, measurement of gate visitors, and the like as a monitoring device for monitoring using image processing. , Very effective.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a monitoring device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a processing procedure of this embodiment.

FIG. 3 is an explanatory diagram of a process of detecting a region where a moving object exists according to the present embodiment.

FIG. 4 is a diagram showing a characteristic example of an evaluation function according to the present embodiment.

FIG. 5 is a diagram illustrating an example of setting a gaze area according to the present embodiment.

FIG. 6 is a block diagram showing a configuration example of a conventional monitoring device.

FIG. 7 is a diagram illustrating an example of setting a gaze area in a conventional device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 computer 11 CPU 12 image input unit 13 variable gaze area setting unit 14 moving object detection unit 15 moving object tracking unit 16 gaze area setting unit 17 data bus 2 display unit 3 memory 30a to 30c image of moving object to be monitored 31 movement Center of gravity position of area where object exists 32 Circumscribing rectangle of area where moving object exists 33a to 33c Variable gaze area 34 Gaze area 4 Image capturing device

Claims (3)

[Claims] 1. An apparatus for monitoring a specific moving object invading a monitoring target area, the image capturing apparatus capturing an image of the monitoring target area, the captured image being image-processed at a predetermined time interval, and moving. In a monitoring device comprising moving object detection means for detecting an object, and a gaze area setting means for setting a gaze area for the image processing at a detection predicted position of the moving object at the next time, the detected movement A first means for generating history information representing the trajectory of the object based on the position and dimensions of successive past and present time images of the object, and an evaluation value representing the complexity of the past trajectory from this history information. While calculating
Second means for estimating the next time dimension of the moving object based on this evaluation value and the current time dimension of the moving object, and the dimension of the gaze area at the next time according to the estimated dimension And a third means for updating the monitoring device. 2. The monitoring device according to claim 1, further comprising fourth means for tracking the moving object by associating the history information and the detection positions of the moving object at successive current times with each other. Characteristic monitoring device. 3. The monitoring device according to claim 1, wherein the evaluation value is a value obtained by multiplying a change in the curvature of the past trajectory of the moving object and a change in the moving speed by multiplying each by a predetermined coefficient. A monitoring device characterized by the above.