WO2010024281A1

WO2010024281A1 - Monitoring system

Info

Publication number: WO2010024281A1
Application number: PCT/JP2009/064844
Authority: WO
Inventors: 俊和赤間
Original assignee: 有限会社ラムロック映像技術研究所; 株式会社修成工業
Priority date: 2008-08-28
Filing date: 2009-08-26
Publication date: 2010-03-04
Also published as: JPWO2010024281A1; JP5047361B2

Abstract

Information of the upper limit and the lower limit of the shape of an object desirably detected from among images is stored in an object reference information storage means (1) as reference information of the object, and whether the shape of an object is within the upper and lower limits of the reference information stored in the object reference storage means (1) is determined from among inputted images by using an image analyzing means (3). Thus, only the desired shape is to be a target and manpower cost and labor of an observer can be saved.

Description

Monitoring system

The present invention relates to a monitoring system. More specifically, the present invention relates to a monitoring system that monitors the actions of suspicious persons or persons who want to watch over, for the purpose of crime prevention or watching.

Conventional monitoring systems generally have a function of simply recording a video by connecting a video signal from a camera to a monitor and a recording device such as a VTR.
For this reason, in order to prevent crimes in advance, there are problems in terms of labor costs and labor of the person (monitorer) to be monitored, such as the fact that the person must always be watching the appearance of the person being shown.

In recent years, cameras (hereinafter referred to as “network cameras”) that can watch images from remote locations using an Internet line or the like are becoming widespread, and using them reduces the time and effort required to check the images on site. be able to.
However, even in the case of such a network camera, it is necessary to always watch the camera image, and using the network camera alone cannot improve the problems in terms of labor cost and labor of the observer.

In addition, the surveillance system has an increasing social need not only for the purpose of crime prevention described above but also in terms of “watching”.
For example, there is a problem that elderly people living alone (hereinafter referred to as “old people living alone”) die lonely. It is necessary to be resident or to install a network camera at the site and always monitor the video transmitted from the network camera by someone from a remote place.
In other words, “watching” also has a problem in terms of labor cost and labor when a conventional monitoring system is used.

By the way, as a technique to reduce the above problems, in order to detect the presence of suspicious persons and criminals, a thermal sensor that reacts to people, etc., is installed in addition to the camera, and an Internet line etc. For example, when a thermal sensor detects a person, a method of notifying that a person has been detected by sending an e-mail, for example, to a remote monitor via the Internet line is conceivable.
However, this method has a problem that a suspicious person cannot be detected at a location where a heat sensor cannot be installed (for example, a public road in front of the house).

In addition, for example, in order to watch the life of the elderly living alone, it is conceivable to install a network camera and a heat sensor inside the elderly living alone and provide an internet line.
However, because the thermal sensor reacts not only to the suspicious person's intrusion but also to the elderly living there, every time a remote supervisor is notified, the supervisor frequently views the scene image. It will be necessary to check visually, and the labor of the supervisor cannot be reduced.

In the above, there is also a method using a moving object detection sensor instead of the thermal sensor. The moving object detection sensor is a technique for analyzing a video signal from a camera and detecting a screen change. In this case, a heat sensor is unnecessary.
A specific method of the moving object detection sensor stores a plurality of videos (frames) at different times. For example, the current frame and the previous frame are stored. Then, the absolute value of the difference between the current frame and the immediately preceding frame is obtained as the amount of change, and when the amount of change is greater than a certain threshold, it is determined that there is movement.

However, these conventional monitoring systems have various problems described below.

First, the first problem in the conventional monitoring system is that if there are moving objects or subjects, the thermal sensor or moving object detection sensor reacts to small animals such as dogs and cats and determines that there is an abnormality. Is a point.
For example, when used for crime prevention outdoors, it reacts not only to intruders but also to small animals such as dogs and cats. In addition, for example, if a cat or a motion detection sensor reacts every time a cat passes through the site and notifies the observer that there is an abnormality via the Internet line, there is no possibility of a crime. Therefore, the supervisor is busy with the situation confirmation work such as checking the video of the network camera and dispatching people to the scene to confirm the safety situation of the scene. What is a good monitoring system? I can not say.

A second problem in the conventional monitoring system is that, if there is a moving object or a subject, the moving body detection sensor reacts to an object other than a person and determines that there is an abnormality.
For example, when used for crime prevention outdoors, not only intruders but also cars that are traveling on the road in front of the site will react. In addition, for example, when the vehicle detection sensor reacts every time a car passes the road and notifies the observer that there is an abnormality through the Internet, the observer In order to check the safety status of the camera, it is not possible to say that it is a good monitoring system because it will be followed by the status check work such as checking the video of the network camera or dispatching people to the site.

In addition, the third problem in the conventional monitoring system is that, for example, even if a heat sensor or a motion detection sensor is installed in the living room of an elderly living alone to watch the life of the elderly living alone, for example, Even if it gets worse, it's crouching or falling down, it can't distinguish between the normal state and the situation, and someone stays on site, frequently visits on site, or network camera It is necessary to keep watching the video, and there are still problems in terms of labor costs and labor of the observer.

The present invention has been made in view of the above points, and an object thereof is to provide a monitoring system capable of solving the first to third problems.

In order to achieve the above object, in the monitoring system of the present invention, object reference information storage means for storing information on the upper and lower limits of the shape of the object is input as reference information of the object to be detected from the video. Video analysis means for determining whether or not the shape of the object in the video is within the upper and lower limits of the reference information stored in the object reference information storage means.

Here, as reference information of the object to be detected from the video, information on the upper limit (for example, large rectangle) and the lower limit (for example, small rectangle) of the shape of the object is stored in the object reference information storage means, and the video analysis means is By determining whether or not the shape of the object in the input video is within the upper and lower limits of the reference information stored in the object reference information storage means, the first and second problems described above Can be solved.

Specifically, for example, when only people such as intruders are to be detected, the large rectangle is made slightly larger than the size of the person, and the small rectangle is made larger than the small animal. Since a larger automobile or the like can be excluded from the detection target, the first problem and the second problem can be solved as described above.

In addition, for example, by setting a large rectangle and a small rectangle corresponding to a shape and size corresponding to a state where a person has fallen, it is possible to detect a state in which an elderly person falls in a bad condition. And the above third problem can be solved.

In order to achieve the above object, in the monitoring system of the present invention, object reference information storage means for storing upper and lower limit information of the shape of an object as reference information of an object to be detected from an image, and imaging Effective area information storage means for storing, as effective area information, information specifying a predetermined area in the area to be recorded, and an object in the input video in the effective area information stored in the effective area information storage means Image analysis means for determining whether or not the object is located and determining whether the shape of the object is within the upper and lower limits of the reference information stored in the object reference information storage means.

Here, information specifying a predetermined area in the imaged area is stored as effective area information in the effective area information storage means, and an object in the video input using the video analysis means is stored in the effective area information storage means. By determining whether or not the position is within the stored effective area information, the first problem can be solved with higher accuracy.

Specifically, for example, a distant human and a nearby bird are actually the same size on the screen, even though they are completely different in size. There is a possibility that. In such a case, for example, by limiting the effective area information only to a distant place, the problem of detecting a nearby bird can be avoided, and the above-mentioned first purpose is to distinguish small animals from humans. The first problem can be solved with higher accuracy.

The situation of the installation site of an imaging device such as a camera is provided by providing video display means for displaying video input to the video analysis means and object reference information input means for inputting reference information to the object reference information storage means. The first, second, and third problems can be solved without being affected by the above.

Specifically, for example, a television monitor is used as an image display means for displaying an image, and a mouse used in a computer as an object reference information input means for inputting reference information of an object to be detected stored in the object reference information storage means. For example, by observing the shape and size of a person appearing on the screen of a television monitor, for example, by using a mouse to specify a rectangle that is close to the shape and size of the person, an object ( For example, even if the shape or size of a person changes, the object (person) can be detected, and the first problem, the second problem, and The third problem can be solved.

Similarly, an installation site of an imaging device such as a camera is provided by including video display means for displaying video input to the video analysis means and effective area information input means for inputting effective area information to the effective area information storage means. The first problem, the second problem, and the third problem can be solved without being affected by the situation.

Specifically, for example, a mouse used in a computer as an effective area information input means for inputting effective area information of an object to be detected, which is stored in an effective area information storage means, is used as a video display means for displaying an image. By using the mouse to specify the effective area information while visually observing the shape and size of the person appearing on the screen of the TV monitor, it is not affected by the situation of the camera installation site. The first problem, the second problem, and the third problem can be solved.

Also, the analysis result notification means for notifying the judgment result of the video analysis means can notify the monitor of the situation.
For example, by using communication means using the Internet as analysis result notification means, it becomes possible to notify the situation to a remote supervisor. Further, for example, by using an acoustic means using a speaker or the like as the analysis result notification means, the situation can be notified to the supervisor using voice. Furthermore, by using a light emitting means using a lamp or the like as the analysis result notification means, the situation can be notified to the supervisor using the lighting of the lamp.

In addition, in the case of including a video transmission unit configured to transmit a video input to the video analysis unit to a terminal connected via a communication line, the video transmission unit Privacy can be protected by transmitting video.
The “communication line” here means a path for signal transmission / reception, and includes both a wired case and a wireless case. “Terminals connected via a communication line” "Includes a terminal connected by a wired cable such as a LAN cable, a terminal connected by a wireless LAN, a mobile phone terminal, and the like.

For example, when transmitting the video of the monitoring system to a remote terminal, the video from the monitoring system is not always transmitted, but the video from the monitoring system is transmitted only in the event of an emergency, by the monitoring system The privacy of the person being monitored can be protected.
Even if the video from the surveillance system is always transmitted, it is possible to display the video received on the terminal side only when the signal indicating the emergency state is transmitted together with the signal indicating whether the emergency state is received or not. Privacy protection is also realized by configuring.

In order to achieve the above object, the monitoring system of the present invention stores the upper and lower limit information of the shape of the first object as reference information of the first object to be detected from the video, Object reference information storage means for storing upper and lower limit information of the shape of the second object as reference information of the second object to be detected from the image, and the shape of the object in the input image is the object First image analysis means for determining whether the reference information of the first object stored in the reference information storage means is within the upper and lower limits; and the shape of the object in the input image is the object reference A second video analysis unit that determines whether the reference information of the second object stored in the information storage unit is within the upper and lower limits; a determination result of at least the first video analysis unit; and the second Logic based on the results of video analysis And a logic determination means for performing constant.

Here, at least based on the determination result of the first video analysis means and the determination result of the second video analysis means, the logical determination (for example, the determination result of the first video analysis means and the determination result of the second video analysis means). The above third problem can be solved by the logic determination means that performs logic determination based on AND and NAND.

Specifically, for example, from the video signals of two cameras, the first camera detects a fallen person, and the second camera detects a pedestrian different from the fallen person. Usually, when the first camera detects a fallen person, notification is required to call for help, but when the second camera detects another pedestrian, a pedestrian (that is, a person who can assist) ), And there is no need to call for new help. Therefore, when the second camera detects a pedestrian in the logic determination means, the detection of the faller of the first camera is NANDed (denied), so that the supervisor's labor can be reduced by not using help. Therefore, the third problem can be solved.

“At least based on the judgment result of the first video analysis means and the judgment result of the second video analysis means” means “the judgment result of the first video analysis means and the judgment result of the second video analysis means”. In addition to the “judgment result of the first video analysis unit and the judgment result of the second video analysis unit”, in addition to the case where the logical determination is performed based on “ This also includes a case where a logical determination is made based on the determination result.

In order to achieve the above object, in the monitoring system of the present invention, object reference information storage means for storing upper and lower limit information of the shape of an object as reference information of an object to be detected from an image, and input Video analysis means for determining whether the shape of the object in the received video is within the upper and lower limits of the reference information stored in the object reference information storage means, and a terminal connected via a communication line On the other hand, a first monitoring system, a second monitoring system, and a video transmission means of at least the first monitoring system, each having video transmission means configured to transmit the video input to the video analysis means, and Storage means for readable storage of the reception history information of the video together with the video transmitted from the video transmission means of the second monitoring system, and the reception history information stored in the storage means It has Shimesuru display means and selecting the reception history information displayed on the display unit, and a terminal for video corresponding to the received history information is configured to be displayed on the display means.

In order to achieve the above object, in the monitoring system of the present invention, object reference information storage means for storing upper and lower limit information of the shape of an object as reference information of an object to be detected from an image, and imaging Effective area information storage means for storing, as effective area information, information specifying a predetermined area in the area to be recorded, and an object in the input video in the effective area information stored in the effective area information storage means A video analysis means for judging whether or not the position of the object is within the upper and lower limits of the reference information stored in the object reference information storage means, and a communication line A first monitoring system, a second monitoring system, and at least the first monitoring system, each having video transmission means configured to transmit the video input to the video analysis means to a connected terminal; Storage means for readablely storing the reception history information of the video together with the video transmitted from the video transmission means of the monitoring system and the video transmission means of the second monitoring system, and the reception history information stored in the storage means And a terminal configured to display a video corresponding to the reception history information on the display means by selecting the reception history information displayed on the display means.

Here, by selecting the reception history information displayed on the display means, the video corresponding to the reception history information is configured to be displayed on the display means, so that the images are simultaneously received from a plurality of surveillance cameras. In addition, a desired video can be confirmed based on the reception history information.

The “communication line” here means a path for signal transmission / reception, and includes both a wired case and a wireless case. “Terminals connected via a communication line” "Includes a terminal connected by a wired cable such as a LAN cable, a terminal connected by a wireless LAN, a mobile phone terminal, and the like.

In addition, “at least the video transmitted from the video transmission unit of the first monitoring system and the video unit of the second monitoring system” means “the video transmission unit of the first monitoring system and the video unit of the second monitoring system”. In addition to the “video transmitted from the video transmission means”, in addition to “video transmitted from the video transmission means of the first monitoring system and the video transmission means of the second monitoring system”, 3 ”is also included.

The monitoring system of the present invention can solve the first to third problems described above.

It is the figure which showed the structure of Example 1 of the monitoring system of this invention. It is the figure which showed the state when a camera image | photographed a person from the top. It is the figure which showed the state when a camera image | photographed the person from the side horizontally. It is the figure (1) which showed the state of the small rectangle and the large rectangle. It is FIG. (2) which showed the state of the small rectangle and the large rectangle. It is the figure which showed the vertical length and horizontal length of an object (person). It is the figure which showed the shape and size of a person and a car. It is the figure which showed the state of the absolute value of the difference in case a person does not exist in both the past frame and the present frame. It is the figure which showed the state of the absolute value of the difference in case a person exists in the present flame | frame. It is the figure which showed the state which the camera is image | photographing from the side that a person is falling down. It is the figure which showed the state in which the person in the image | video of a camera has fallen. It is the figure which showed the state which the camera is image | photographing from the side that a person is falling down. It is the figure which showed the state in which the person in the image | video of a camera has fallen. It is the figure which showed the example of the rectangular set for detecting the person who has fallen. It is the figure which showed the example of the rectangular set for detecting the person who has fallen. It is the figure which showed the state which the camera is image | photographing the person who is standing from the top. It is the figure which showed the camera image | video when the camera image | photographs the person who is standing from the top. It is the figure which showed the state which has image | photographed the person who has fallen from the top. It is the figure which showed the camera image | video when a camera image | photographs the person who has fallen from the top. It is the figure which showed the example of the rectangular set for detecting the person who has fallen. It is the figure which showed the example of object reference | standard information other than a rectangle. It is the figure which showed the state of the person far from a camera and the state of a bird far away. It is the figure which showed the state of the person far from a camera, and the nearby bird. It is the figure which showed the example of the effective area rectangle which applies object reference information. It is the figure which showed the example which has photographed the fallen person with the camera 1, and another pedestrian with the camera. It is the figure which showed the structure of Example 2 of the monitoring system of this invention. It is the figure which showed the structure of Example 3 of the monitoring system of this invention.

Hereinafter, modes for carrying out the present invention (hereinafter referred to as “examples”) will be described with reference to the drawings.

A first embodiment of the present invention for solving the problems of the conventional monitoring system will be described below.

FIG. 1 shows the configuration of a monitoring system according to the first embodiment of the present invention.
In FIG. 1, reference numeral 1 denotes a video storage means for storing a video signal, and as an example, it is assumed that videos of a plurality of frames at different times are stored. Reference numeral 2 denotes an object reference information storage means for storing reference information of an object to be detected from the subject in the video, and 3 denotes an analysis of the video signal stored in the video storage means 1 so that the subject is stored in the object reference information. An image analysis means for determining whether or not the object reference information stored in the means 2 matches, 4 is an analysis result information storage means for storing the analysis result information of the image analysis means 3, 5 is a computer, Assuming that the video storage means 1, the object reference information storage means 2, the video analysis means 3, and the analysis result information storage means 4 are implemented as software programs on the computer 5. Reference numeral 6 denotes a camera, and reference numeral 7 denotes a monitor (video display means) for displaying video. A mouse 8 is used as an object reference information input unit. Assume that the camera 6 can convert an analog video signal into a digital signal and input it to the computer 5.

First, an outline of the operation of this monitoring system will be described.
For example, in the case of use for crime prevention outdoors, there may be a utilization method in which an intruder (person) is detected and, for example, an alarm is sounded or a remote supervisor is notified.
However, not only humans but also small animals such as dogs and cats exist outdoors. When a warning is sounded every time a dog or cat passes or a notification is sent to a remote location, the monitor will increase the work of checking the video on the site, so only in case of an important event such as an intruder (person) appearing A system in which notification is performed is desired.

In order to realize this, in this embodiment, the video signal of the camera 6 is stored in the video storage means 1 and the stored video is analyzed by the video analysis means 3 to determine the shape and size of the subject. . Information on the shape and size of the intruder (person) to be detected in advance is stored in the object reference information storage means 2 as object reference information, and the object reference information is compared with the shape and size of the subject. Are determined, that is, whether the shape and size of the intruder (person) are satisfied, and the result is stored in the analysis result information storage means 4.

Next, the details of the operation of this monitoring system will be described.

No reaction to small animals such as dogs and cats (no alarms or notifications to remote locations), but to detect and respond only to intruders (people) (alarms and notifications to remote locations) Need to automatically distinguish between small animals and people. For this purpose, it is necessary to determine the shape and size of the subject in the image of the camera 6 in FIG.

However, since the shape and size of the subject in the video changes depending on the distance and angle between the camera and the subject, for example, when “object” is to be detected, information on the shape and size (object reference information) There is a problem that it is not possible to cope with a fixed one.

Specifically, the case where the camera 6 has taken a person from the top as shown in FIG. 2A and the case where the camera 6 has taken a person from the side as shown in FIG. 3A. Then, even if the same person is photographed, the shape looks very different. The camera installation position is not always horizontal as shown in FIG. 3A, and there are infinite combinations of distances and angles between the camera and the subject depending on the situation. .

FIG. 2B schematically shows a camera image when the camera is photographed from the top to the bottom as shown in FIG. 2A, and FIG. As shown in a), a camera image in a case where the camera photographs a person from the side to the side is schematically shown.

Thus, in order to flexibly cope with the problem that the shape and size of the subject changes depending on the installation position of the camera, it has a function of inputting object reference information (shape and size) of the object to be detected. This is one of the important features of the present invention.

In the monitoring system of the present invention, first, information (object reference information) on the shape and size of an object (eg, a person) to be detected is stored in the object reference information storage means 2.

As an example of the method for inputting the object reference information, a small rectangle (hereinafter referred to as “small rectangle”) is used as the lower limit of the object reference information, as shown in FIG. B) and a large rectangle (hereinafter referred to as “large rectangle”) A as an upper limit.
Here, the small rectangle in FIG. 4 is set to be slightly larger than the small animal (cat) b. The large rectangle shown in FIG. 4 is set to be slightly larger than the size of the person a. For example, the vertical length and the horizontal length of each of the small rectangle and the large rectangle are stored in the object reference information storage unit 2 in FIG. 1 as reference information of the object to be detected.

When the small rectangle B and the large rectangle A are set as described above, as shown in FIG. 5, an object larger than the small rectangle B and smaller than the large rectangle A (the size and shape corresponding to the hatched portion in FIG. Object).

Specifically, FIG. 6 is a diagram showing the vertical length D and the horizontal length E of the object (person) C. This vertical length D is the vertical length of the small rectangle in FIG. 6 and below the vertical length of the large rectangle, and the horizontal length E in FIG. 6 is not less than the horizontal length of the small rectangle in FIG. 5 and not more than the horizontal length of the large rectangle. If there is, such an object (person) will be detected.

By defining the small rectangle B and the large rectangle A in this way, a small animal such as a cat smaller than the small rectangle as shown in FIG. 4 is excluded from the detection target. A “person” smaller than the large rectangle becomes a detection target, and an excellent monitoring system that does not react to small animals such as dogs and cats can be constructed.

Here, it is important to set not only a small rectangle but also a large rectangle.
For example, as shown in FIG. 7, a situation is assumed in which an automobile e or the like sufficiently larger than the person d is shown in the video. Specifically, it is a case where an automobile passes through a road in front of the site outdoors. In that case, the car simply passes through the road, and there is no particular problem in terms of crime prevention. However, if the car reacts each time it passes (alarms or notification to remote locations), This leads to problems that cannot be reduced. Therefore, by setting a large rectangle and ignoring automobiles that are larger than the large rectangle, it is possible to detect "people" smaller than the large rectangle, for example, trying to enter a building or site. It reacts only to "people" (intimidating with alarms or notifying remote observers).
And since this makes it possible to construct an efficient monitoring system that places little burden on the observer, it is important to set not only the small rectangle but also the large rectangle as described above.

Next, in the monitoring system of the present invention, the video signal from the camera 6 in FIG. 1 is stored in the video storage means 1 and the video analysis means 3 analyzes the shape and size of the subject.
The analysis method is not limited. For example, there is a method of obtaining an absolute value of the interframe difference between a past frame several seconds before where no subject (moving object) is reflected and a current frame where the subject exists. Can be mentioned.

Specifically, FIG. 8 (FIG. 8A shows the past frame, FIG. 8B shows the current frame, and FIG. 8C shows the absolute value of the difference.) As described above, when only the background is shown in the past frame and only the background is shown in the current frame, the absolute value of the difference between the past frame and the current frame becomes small. However, as shown in FIG. 9 (FIG. 9A shows the past frame, FIG. 9B shows the current frame, and FIG. 9C shows the absolute value of the difference). The past frame is only the background, but when the person f is shown in the current frame, the absolute value of the difference between the past frame and the current frame becomes large. That is, the absolute value of the difference in the portion where the person is shown increases.
Therefore, it is possible to know the shape and size of the subject by extracting a portion where the absolute value of the difference is large.

In this embodiment, the shape and size information (object reference information) of the “person” to be detected stored in the object reference information storage unit 2 in FIG. 1 is compared with the shape and size of the subject. The result is stored in the analysis result information storage unit 4.

It should be noted that the result stored in the analysis result information storage means 4 is “intruder (person) by connecting with the external system according to the purpose, such as connecting the monitoring system of this embodiment and the Internet line”. If the result is `` Yes '', for example, a remote person is notified by e-mail, a warning sound is emitted through an external speaker, or an external light is emitted, depending on the purpose. Various applications are possible.

Here, by linking the monitoring system of this embodiment with an external system, the monitoring system video is transmitted to a remote terminal (for example, a stationary personal computer or a portable small personal computer), and Although it is possible to check the video on the terminal, in consideration of privacy protection issues, an emergency situation occurs, such as the video being hidden in the normal state (the state where no abnormality has occurred) and the person falling down. It is preferable that the video (moving image or still image) is displayed only in the case where it is performed.

In addition, when a plurality of (for example, two) monitoring systems of this embodiment are used to watch over a plurality of places, a plurality of cameras transmit images of occurrence of an abnormality simultaneously or one after another. In such a case, it may be possible that the supervisor is overlooked, for example, by switching the first video to the later video. For this reason, it is preferable that the history is displayed on the terminal of the supervisor and the video selected from the history can be viewed.

In the monitoring system of the present embodiment, by specifying a rectangular set (meaning a combination of a small rectangle and a large rectangle stored in the object reference information storage unit 2 in FIG. 1), small animals such as dogs and cats, You can ignore the car passing. Therefore, every time a small animal or car passes, the supervisor does not have to fend for on-site confirmation, and it is only necessary to respond when there is an intruder (person), creating an efficient surveillance system can do.

Moreover, in the monitoring system of a present Example, it can respond to various uses by changing the setting of a rectangular set.
For example, it can be used to distinguish between children and adults. In that case, a small rectangle that is slightly larger than an infant or a child is set, and a large rectangle that is slightly larger than an adult person is set. By setting in this way, for example, it can be applied as a system that automatically notifies the staff room when an adult (suspicious person) comes in without reacting to an infant or child in a kindergarten or elementary school. it can.

In the case of conventional thermal sensors and motion detection sensors that do not have such functions, notifications are generated not only for adults but also for children (infants and children), so notifications that are not important for crime prevention Will occur frequently, and each notification must be confirmed by the person, making it impossible to reduce the labor of the person who is monitoring, May be overlooked when a suspicious person is invading. On the other hand, such a problem can be greatly reduced in the monitoring system of the present embodiment.

In the monitoring system of the present embodiment, it is possible to flexibly cope with the problem that the subject looks different depending on the installation position of the camera.
For example, as shown in FIG. 3A, when the camera captures a person from the side horizontally, the person (subject) in the video looks like a long rectangle. On the other hand, for example, as shown in FIG. 2A, when the camera captures a person from above, the image of the camera looks circular or nearly square. In this way, even the same person (subject) looks different depending on the relative relationship with the position of the camera to shoot. In view of these circumstances, the monitoring system of the present embodiment has a system that is less dependent on the installation location because the shape of the object to be detected can be specified according to the site while viewing the screen while the camera is installed. can do.

In the monitoring system of the present embodiment, it is also possible to detect “a state where a person has fallen” by setting a plurality of rectangular sets.
For example, an elderly person living alone might fall seriously and fall down, leading to a serious accident that would result in death if left unattended for several days. It is possible to prevent such a serious accident of an elderly person living alone by detecting “a state where a person has fallen”.

However, the “state where a person has fallen” appears to be a different shape and size as a camera image, depending on the relative relationship between the position of the camera and the position of the person who has fallen. For example, as shown in FIG. 10, when the camera is photographed from the side in a state where a person is lying down, the video of the camera looks as shown in FIG. In addition, when the camera photographs a person who is lying down as shown in FIG. 12 from the side, the video of the camera looks as shown in FIG. As can be seen by comparing FIG. 11 and FIG. 13, even if the camera captures a person who has fallen from the same position, it looks so different depending on the direction of the fall.

Even in these cases, the monitoring system of the present embodiment can sufficiently cope. That is, the object reference information storage means shown in FIG. 1 is designated as reference information of an object to be detected by designating a plurality of rectangular sets so that both of FIGS. 11 and 13 can be detected as being in a fallen state. 2 and when the subject that matches any of the plurality of rectangular sets is detected in the video analysis means 3, it is determined that “the person has fallen”, thereby sufficiently indicating “the person has fallen” It becomes possible to detect.

Specifically, a rectangle set as shown in FIG. 14 is designated to detect the fallen state as shown in FIG. 11, and a rectangle as shown in FIG. 15 is used to detect the fallen state as shown in FIG. Specify a set. In addition, by specifying a plurality of rectangular sets according to how to fall, it is possible to improve the accuracy of detecting a state where a person is falling. In FIG. 14 and FIG. 15, symbol A indicates a large rectangle, and symbol B indicates a small rectangle.

In addition, for example, when the camera 6 is installed so as to photograph a person from above as shown in FIG. 16, a standing person looks like FIG. 17 in the image of the camera. However, when a person who is lying down is photographed from above as shown in FIG. 18, the video of the camera looks as shown in FIG. The image of the camera in FIG. 19 when the camera captures a person from above is different from the image of the camera when the camera captures a person from the side (FIGS. 11 and 13). Therefore, a rectangular set as shown in FIG. 20 is required to deal with a fallen image as shown in FIG. As described above, by specifying a plurality of rectangular sets, it is possible to realize a system that can flexibly cope with a change in the installation position of the camera and detect a “person who has fallen”. In FIG. 20, symbol A indicates a large rectangle, and symbol B indicates a small rectangle.

In addition, as an example, the object reference information is represented by a rectangle. However, the object reference information is not limited to a rectangle. For example, as shown in FIG. You may select and use what you want to use, or adjust (enlarge, reduce, or deform) the shape and size of the selected model.

In addition, what you want to detect changes depending on the application, so the model is not limited to the human form, but by preparing various object models such as animals and cars and using them as object reference information, for example, the number of cars It can also be used as a system that can aggregate and count the number of people.

Note that the video stored in the video storage unit 1 of FIG. 1 is not limited to a frame, and may be a field. For example, when converted into luminance and color difference, only luminance or color difference is used. It may be configured, and it may be enlarged or reduced, partly excerpted, frequency converted, filtering processing such as differentiation in the time axis direction or spatial direction, color number change, quantization, etc. Thus, the tone change of each signal may be performed.

In the present embodiment, the camera 6 in FIG. 1 can convert an analog video signal into a digital signal and input it to the computer 5. However, if the camera 6 outputs an analog signal, It is good also as a structure which equips the computer 5 with the function which converts into a digital signal in the computer 6, and if the video signal of the camera 6 is some compression, between the computer 5 and the camera 6, It is good also as a structure provided with a means to expand | extract the signal compressed in the computer 5. FIG.

In this embodiment, the video signal is used while being directly input from the camera 6 to the computer 5. However, the video of the camera 6 may be transmitted to the remote computer 5 using an Internet line or the like. It has the same meaning.

Note that the method of extracting the shape and size of the subject of the camera 6 is not limited to the method shown in this embodiment.

In this embodiment, the main configuration of the monitoring system is software that operates on a computer. In that case, the monitoring system is in the form of a program that operates on the processor of the computer. Various controls are performed by the CPU, and various storage means. Consists of a computer memory and hard disk. These functions may be configured as system LSI, other hardware, software operating on a computer, hardware incorporated in a computer, or both software and hardware. Furthermore, the recording medium which recorded the program which operate | moves on a computer may be sufficient.

In addition, although the case where the number of cameras was one was shown in the present Example, there may be a plurality of cameras.

Further, not only those that receive visible light such as a color camera, but those that can receive any optical signal, such as an infrared light receiver, can substantially acquire an image, so these are also described in the present embodiment. Included in the camera.

Hereinafter, a second embodiment of the present invention for solving the problems of the conventional monitoring system will be described.

FIG. 26 shows the configuration of the monitoring system according to the second embodiment of the present invention.
In FIG. 26, reference numeral 1 denotes video storage means for storing a video signal, and it is assumed that videos of a plurality of frames having different times are stored as an example. Reference numeral 2 denotes an object reference information storage means for storing reference information of an object to be detected from the subject in the video, and 3 denotes an analysis of the video signal stored in the video storage means 1 so that the subject is stored in the object reference information. An image analysis means for determining whether or not the object reference information stored in the means 2 matches, 4 is an analysis result information storage means for storing the analysis result information of the image analysis means 3, 5 is a computer, Assuming that the video storage means 1, the object reference information storage means 2, the video analysis means 3, and the analysis result information storage means 4 are implemented as software programs on the computer 5. Reference numeral 6 denotes a camera, and reference numeral 7 denotes a monitor (video display means) for displaying video. A mouse 8 is used as an object reference information input unit. The camera 6 can convert an analog video signal to digital and input it to the computer 5, and the above is the same as the configuration of FIG. Reference numeral 9 denotes effective area information storage means for storing effective area information by using effective area coordinates in the screen of the reference information of the object as effective area information. 8 mouse is also used as an effective area information input means.
The difference from the first embodiment is that this effective area information storage means 9 is provided.

In order to realize this, in this embodiment, the video signal of the camera 6 is stored in the video storage means 1 and the stored video is analyzed by the video analysis means 3 to determine the shape and size of the subject. . Information on the shape and size of the intruder (person) to be detected in advance is stored in the object reference information storage means 2 as object reference information, and the object reference information is compared with the shape and size of the subject. Are determined, that is, whether the shape and size of the intruder (person) are satisfied, and the result is stored in the analysis result information storage means 4.
Up to this point, the operation is the same as in the first embodiment.

This embodiment is different from the first embodiment in that the effective area coordinates to which the object reference information is applied are stored in the effective area information storage means 9 in the screen, and the effective area is limited so that a bird or the like This is to improve the accuracy of discriminating between small animals and humans.

FIG. 22 shows a situation where a person g and a bird h exist in the distance. Both humans and birds exist at the same distance from the camera. In this case, the size of each is completely different, so by setting a small rectangle sufficiently larger than the bird as object reference information, Is not mistaken for a person.

However, the case as shown in FIG. 23 has a problem. That is, FIG. 23 shows a state in which a person g exists in the distance and a bird h exists in the vicinity of the camera. As for the size, the case where a bird appears as large as a human being is a problem. In this case, since the difference in size and shape is small, there is a possibility that a bird is mistakenly recognized as a person.
In order to suppress such a problem, in this embodiment, the coordinates of the area to which the object reference information is applied are set as the effective area rectangle X as shown in FIG. It is possible to avoid misrecognizing that a bird existing nearby is a person.

Next, the details of the operation of this monitoring system will be described.

No reaction to small animals such as dogs and cats (no alarms or notifications to remote locations), only reacting by detecting intruders (people) (alarms or notifications to remote locations) If you want to, you need to automatically distinguish between small animals and people. For this purpose, it is necessary to determine the shape and size of the subject in the video of the camera 6 in FIG.

However, since the shape and size of the subject in the video changes depending on the distance and angle between the camera and the subject, for example, when “object” is to be detected, information on the shape and size (object reference information) There is a problem that it is not possible to cope with a fixed one.
Specifically, the case where the camera 6 has taken a person from the top as shown in FIG. 2A and the case where the camera 6 has taken a person from the side as shown in FIG. 3A. Then, even if the same person is photographed, the shape looks very different. The camera installation position is not always horizontal as shown in FIG. 3A, and there are infinite combinations of distances and angles between the camera and the subject depending on the situation. .

In the monitoring system of the present invention, first, information (object reference information) on the shape and size of an object (for example, a person) to be detected is stored in the object reference information storage means 2.
As an example of a method for inputting the object reference information, a small rectangle B as a lower limit and a large rectangle A as an upper limit of the object reference information as shown in FIG. Set.
Here, the small rectangle in FIG. 4 is set to be slightly larger than the small animal (cat) b. The large rectangle shown in FIG. 4 is set to be slightly larger than the size of the person a. For example, the vertical and horizontal lengths of the small rectangle and the large rectangle are stored in the object reference information storage unit 2 in FIG. 26 as reference information of the object to be detected.

When the small rectangle B and the large rectangle A are set as described above, an object larger than the small rectangle B and smaller than the large rectangle A (a size and a shape corresponding to the shaded portion in FIG. 5), as shown in FIG. An object having a) is detected.

Specifically, FIG. 6 is a diagram showing the vertical length D and the horizontal length E of the object (person) C. This vertical length D is the vertical length of the small rectangle in FIG. 6 and below the vertical length of the large rectangle, and the horizontal length E in FIG. 6 is not less than the horizontal length of the small rectangle in FIG. 5 and not more than the horizontal length of the large rectangle. If there is, such an object (person) is detected.

Next, in the monitoring system of the present invention, the video signal from the camera 6 of FIG. 26 is stored in the video storage means 1 and the video analysis means 3 analyzes the shape and size of the subject.
The analysis method is not limited. For example, there is a method of obtaining an absolute value of the interframe difference between a past frame several seconds before where no subject (moving object) is reflected and a current frame where the subject exists. Can be mentioned.

Specifically, as shown in FIG. 8, when only the background is shown in the past frame and only the background is shown in the current frame, the absolute value of the difference between the past frame and the current frame becomes small. However, as shown in FIG. 9, the past frame is only the background, but when the person f is reflected in the current frame, the absolute value of the difference between the past frame and the current frame becomes large. That is, the absolute value of the difference in the portion where the person is shown increases.
Therefore, it is possible to know the shape and size of the subject by extracting a portion where the absolute value of the difference is large.

In this embodiment, an effective area rectangle is set as shown in FIG. 24, for example, using the mouse 8 while watching the image on the monitor 7 in FIG. For example, the coordinate information of the upper left end and the lower right end of this effective area is stored as effective area information in the effective area information storage means 9 of FIG.

In the monitoring system of the present invention, the shape and size information (object reference information) of the “person” to be detected stored in the object reference information storage unit 2 in FIG. 26 is compared with the shape and size of the subject. And whether or not the subject exists in the effective area stored in the effective area information storage means 9 and the result is stored in the analysis result information storage means 4. To do. By providing an effective area to which the object reference information is applied in this way, only a distant person within the effective area is detected as shown in FIG. 24, and a bird near the effective area is erroneously recognized as a person. That can be suppressed.

It should be noted that the result stored in the analysis result information storage means 4 is “there is an intruder (person)” by linking with the external system according to the purpose, such as connecting the monitoring system of the present invention and the Internet line. ”, For example, notification to a person at a remote location by e-mail, warning sound through an external speaker, emission of light from an external light, etc. Application becomes possible.

In addition, in this embodiment, a plurality of different rectangular sets are set, a person can be detected in one effective area on the screen, and a car can be detected in another effective area. Possible monitoring system can be constructed.

Furthermore, in the present embodiment, by providing an effective area to which the object reference information is applied, the advantage of the configuration of the first embodiment is provided, and only a far person in the effective area is detected as shown in FIG. Thus, it is possible to suppress erroneous recognition of a bird near the outside of the effective area as a person, so that a monitoring system with higher accuracy and fewer false alarms than the first embodiment can be realized.

Hereinafter, a third embodiment of the present invention for solving the problems of the conventional monitoring system will be described.

FIG. 27 shows the configuration of the monitoring system according to the third embodiment of the present invention.
In the second embodiment, the case where there is one camera is shown as an example. However, in this embodiment, two cameras are provided and various units corresponding to the two cameras are provided. In FIG. 27, reference numeral 1 denotes a video storage means for storing a video signal, and as an example, it is assumed that videos of a plurality of frames having different times are stored. Reference numeral 2 denotes an object reference information storage means for storing reference information of an object to be detected from the subject in the video, and 3 denotes an analysis of the video signal stored in the video storage means 1 so that the subject is stored in the object reference information. The video analysis means for determining whether or not the object reference information stored in the means 2 matches, 4 is an analysis result information storage means for storing the analysis result information of the video analysis means 3, and 9 is an object reference. The effective area information storage means for storing the effective area information using the coordinates of the effective area in the screen as the effective area information, 6 is the camera, and the above is the means for the camera 6.

27, reference numeral 11 denotes video storage means for storing a video signal, and it is assumed that, for example, videos of a plurality of frames having different times are stored. Reference numeral 12 denotes an object reference information storage means for storing reference information of an object to be detected from the subjects in the video. Reference numeral 13 analyzes the video signal stored in the video storage means 11 to store the object reference information. An image analysis means for determining whether or not the object reference information stored in the means 12 matches, 14 is an analysis result information storage means for storing the analysis result information of the image analysis means 13, and 15 is an object reference. Effective area information storage means for storing effective area information using the coordinates of the effective area in the screen as effective area information, 10 is a camera, and the above is means for the camera 10. Reference numeral 16 denotes logical determination means for performing logical determination such as NAND of the analysis result corresponding to the camera 6 and the analysis result corresponding to the camera 10. This is to determine the importance of the detection state by comprehensively considering the analysis results of a plurality of cameras. It is possible to use AND and NAND properly according to the purpose.

Reference numeral 5 denotes a computer, for example, video storage means 1, video storage means 11, object reference information storage means 2, object reference information storage means 12, video analysis means 3, video analysis means 13, analysis result information storage means 4 and analysis. The result information storage unit 14 and the logic determination unit 16 are implemented as software programs on the computer 5. Reference numeral 7 denotes a monitor (video display means) for displaying video. A mouse 8 is used as an object reference information input unit. The camera 6 and the camera 10 can convert analog video signals into digital signals and input them to the computer 5.

First, an outline of the operation of this monitoring system will be described.
FIG. 25 shows an installation example of each camera, and the role of each camera will be described.
The camera 6 photographs from above and is used for detecting a fallen person. The camera 10 shoots from the side and is used to detect a pedestrian different from the fallen person. Normally, when the camera 6 detects a fallen person, notification is required to call for help. However, when the camera 10 detects another pedestrian, a pedestrian (that is, a person who can assist) ), And there is no need to call for new help. Therefore, when the camera 10 detects a pedestrian in the logic determination unit 16, the detection of a fallen person of the camera 6 is NANDed (denied), so that it is applied to a usage that does not call for help. In cases where help is not required, not giving notifications can reduce the effort required for monitoring the site of the supervisor and dispatching the person to the site.

Next, details of the operation of this monitoring system will be described.
First, information (object reference information) of the shape and size of an object (in this case, “falling person”) to be detected by the camera 6 is stored in the object reference information storage unit 2.
Subsequently, information (object reference information) of the shape and size of the object (in this case, “a pedestrian different from the fallen person”) to be detected by the camera 10 is stored in the object reference information storage unit 12.

As an example of a method for inputting object reference information, a small rectangle is set as the lower limit and a large rectangle is set as the upper limit of the object reference information using the mouse 8 while viewing the video on the monitor 7 of FIG. Similarly, effective area information is stored in the effective area information storage unit 15 as necessary. Hereinafter, the flow until the analysis result is stored in the analysis result information storage unit 14 is the same as that of the second embodiment with respect to the unit having the same name.

In this embodiment, when the camera 10 detects the pedestrian j by the logic determination means 16, the detection of the fallen person k of the camera 6 is NANDed. That is, when there is a pedestrian who can help a fallen person, it is determined that the situation is not an important situation requiring notification.

If there is a faller and there are no pedestrians who can help by linking with the external system, such as connecting the monitoring system of this embodiment and the Internet line, for example, by e-mail If there is a pedestrian who can notify and assist, the burden on the supervisor can be greatly reduced by not providing notification.

DESCRIPTION OF SYMBOLS 1 Image | video storage means 2 Object reference | standard information storage means 3 Image | video analysis means 4 Analysis result information storage means 5 Computer 6 Camera 7 Monitor 8 Mouse 9 Effective area information storage means 10 Camera 11 Image | video storage means 12 Object reference | standard information storage means 13 Image | video analysis means 14 Analysis result information storage means 15 Effective area information storage means 16 Logic determination means

Claims

Object reference information storage means for storing information on the upper and lower limits of the shape of the object as reference information of the object to be detected from the image;
A monitoring system comprising: video analysis means for determining whether or not the shape of an object in the input video is within the upper and lower limits of the reference information stored in the object reference information storage means.
Object reference information storage means for storing information on the upper and lower limits of the shape of the object as reference information of the object to be detected from the image;
Effective area information storage means for storing, as effective area information, information specifying a predetermined area in the imaged area;
It is determined whether or not an object in the input video is located within the effective area information stored in the effective area information storage means, and the shape of the object is stored in the object reference information storage means A monitoring system comprising: video analysis means for determining whether or not the reference information is within an upper and lower limit.
Video storage means for storing the input video;
The video analysis means determines whether or not the shape of the object in the video stored in the video storage means is within the upper and lower limits of the reference information stored in the object reference information storage means. The monitoring system described in.
Video storage means for storing the input video;
The video analysis means determines whether or not an object in the video stored in the video storage means is located in the effective area information stored in the effective area information storage means, and the object is The monitoring system according to claim 2, wherein it is determined whether or not the reference information stored in the object reference information storage means is within an upper and lower limit.
The monitoring system according to claim 1, 2, 3, or 4, further comprising analysis result information storage means for storing a judgment result of the video analysis means.
The monitoring system according to claim 1, claim 2, claim 3, or claim 4, comprising notification means for notifying the judgment result of the video analysis means.
Video display means for displaying video input to the video analysis means;
The monitoring system according to claim 1, 2, 3, or 4, further comprising an object reference information input unit that inputs the reference information to the object reference information storage unit.
Video display means for displaying video input to the video analysis means;
The monitoring system according to claim 2, further comprising: effective area information input means for inputting the effective area information to the effective area information storage means.
Video transmission means configured to transmit the video input to the video analysis means to a terminal connected via a communication line, the video transmission means is configured to transmit the video according to the determination result of the video analysis means; The monitoring system according to claim 1, wherein the video input to the analysis unit is transmitted.
Information on the upper and lower limits of the shape of the first object is stored as reference information for the first object to be detected from the image, and the second information is used as reference information for the second object to be detected from the image. Object reference information storage means for storing information on the upper and lower limits of the shape of the object;
First image analysis means for determining whether or not the shape of the object in the input image is within the upper and lower limits of the reference information of the first object stored in the object reference information storage means;
Second video analysis means for determining whether the shape of the object in the input video is within the upper and lower limits of the reference information of the second object stored in the object reference information storage means;
A monitoring system comprising: a logic determination unit configured to perform a logic determination based on at least a determination result of the first video analysis unit and a determination result of the second video analysis unit.
Object reference information storage means for storing upper and lower limit information on the shape of an object as reference information for an object to be detected from the image, and the object shape in the input image is stored in the object reference information storage means. The video analysis means for determining whether the reference information is within the upper and lower limits, and the video input to the video analysis means can be transmitted to a terminal connected via a communication line A first monitoring system and a second monitoring system each having video transmission means;
Storage means for readablely storing the reception history information of the video together with the video transmitted from at least the video transmission means of the first monitoring system and the video transmission means of the second monitoring system; and stored in the storage means Display means for displaying the reception history information, and by selecting the reception history information displayed on the display means, an image corresponding to the reception history information can be displayed on the display means. A monitoring system comprising a terminal.
Object reference information storage means for storing information on the upper and lower limits of the shape of an object as reference information for an object to be detected from the video, and information for specifying a predetermined area in the imaged area is stored as effective area information. The effective area information storage means and whether or not the object in the input video is located within the effective area information stored in the effective area information storage means, and the shape of the object is the object reference Video analysis means for determining whether or not the reference information stored in the information storage means is within the upper and lower limits, and transmission of the video input to the video analysis means to a terminal connected via a communication line A first monitoring system and a second monitoring system each having a video transmission means configured to be possible;
Storage means for readablely storing the reception history information of the video together with the video transmitted from at least the video transmission means of the first monitoring system and the video transmission means of the second monitoring system; and stored in the storage means Display means for displaying the reception history information, and by selecting the reception history information displayed on the display means, an image corresponding to the reception history information can be displayed on the display means. A monitoring system comprising a terminal.