JP6577627B1 - Video surveillance system and method and imaging apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a video surveillance system and method and an imaging device capable of reducing human labor accompanying incident detection. By using a video analysis function of an indent video management server 200 using artificial intelligence, an incident video which seems to be a problem is automatically detected and notified to a person in charge, so that even a small number of people can dramatically increase. The video of the camera 140 can be monitored. When a plurality of monitoring target field systems 100 are used, incident video data for a learning set is successively accumulated in the video recording unit 211, and the performance of artificial intelligence is effectively improved by performing sequential learning. [Selection] Figure 1

Description

  The present invention relates to a video monitoring system, a method thereof, and an imaging apparatus.

  2. Description of the Related Art There is known a system that monitors a monitoring location by displaying an image transmitted from a camera installed at a monitoring location by an administrator terminal owned by an administrator who manages the monitoring location such as a facility.

JP 2006-115435 A

However, there is a problem in that the burden of human labor is heavy because the administrator identifies the incident by viewing the video data at the monitoring location and reports the incident.
In addition, there is a demand for highly accurate incident detection within a short period of time from system startup based on video data and incident information collected from a wide range.
In addition, there is a request to know in advance when the possibility of an occurrence increases, not after an incident occurs.

The present invention has been made in view of such circumstances, and an object of the present invention is to provide a video monitoring system, a method thereof, and an imaging apparatus that can reduce human labor associated with incident detection.
It is another object of the present invention to provide a video monitoring system, a method thereof, and an imaging apparatus that enable highly accurate incident detection in a short period of time.
It is another object of the present invention to provide a video monitoring system, a method thereof, and an imaging apparatus that make it possible to predict the occurrence of an incident in advance.

In order to solve the above-described problems of the prior art and achieve the above-described object, a video monitoring system according to the present invention includes a recording device including a video recording unit that records video data of a video captured by a camera, and the video Incident attribute management means for associating incident video data of an incident detected based on the data with attribute data of the incident, and for a plurality of incidents, the incident video data and the attribute data associated with the incident attribute management means A server for accumulating; level determining means for determining whether or not the possibility of occurrence of the incident exceeds a certain level based on the video data and the attribute data stored in the server; and The incident video data and the attribute data, and the camera Based on the video data of the video that image, and incident detection means for detecting the incident, if the possibility of the incident occurs determines said level decision means to have exceeded a certain level, it is previously associated with the incident Based on the incident response detected by the incident detection means and the attribute data stored in the server used for detection of the incident, the incident and the attribute data Learning means for generating learning data indicating a relationship, wherein the incident attribute management means associates the attribute data indicating an incident occurrence level input by an administrator who viewed the video data with the incident video data , The incident attribute management means, the proactive processing means The incident occurrence possibility level is specified based on the incident video data after executing the pre-processing, and the pre-processing executed by the pre-response processing unit is performed based on the specified incident occurrence possibility level. The learning data is evaluated and updated, the pre-corresponding processing means further performs the pre-processing using the learning data generated by the learning means, and the incident detecting means is generated by the learning means The incident is detected by further using the learned data.

Preferably, the apparatus includes a plurality of incident attribute management means for processing video data captured by the cameras at different monitoring target locations, and the server receives the incident video data received from the plurality of incident attribute management means and the Accumulate attribute data.

Preferably, based on the incident video data and the attribute data received from the incident attribute management means, the attribute data updated according to the operation of the expert is stored in the server in association with the incident video data. It further has management means.

Preferably, the incident attribute management means generates the attribute data indicating the incident occurrence level numerically based on at least one condition of time, place, event holding, and congestion level related to the incident. .

The video monitoring method of the present invention relates to a recording step of recording video data of video captured by a camera in a video recording unit, and to associate incident video data of an incident detected based on the video data with attribute data of the incident Based on the incident attribute management process, a server that stores the incident video data and the attribute data associated in the incident attribute management process for a plurality of incidents, and the video data and the attribute data stored in the server A level determination step for determining whether or not the possibility of occurrence of the incident exceeds a certain level; the incident video data and the attribute data stored in the server; and video data of video captured by the camera; based on, for detecting the incident incident Detecting in the incident detection step, a pre-response processing step for performing a pre-processing previously associated with the incident when the level determination step determines that the possibility of occurrence of the incident has exceeded a certain level And a learning step of generating learning data indicating a relationship between the incident and the attribute data based on the incident and the attribute data stored in the server used for detecting the incident, the incident The attribute management step associates the attribute data indicating the incident probability level inputted by the administrator who viewed the video data with the incident video data, and the incident attribute management step is the pre-response processing step. The incident based on the incident video data after executing pre-processing Identify the raw potential level, the based on the identified incident likelihood level, and evaluating the pre-processing performed by the proactive process updates the learning data, the proactive process Performs the pre-processing using the learning data generated in the learning step, and the incident detection step detects the incident further using the learning data generated in the learning step .

ADVANTAGE OF THE INVENTION According to this invention, the video surveillance system which can reduce the human effort accompanying incident detection, its method, and an imaging device can be provided.
In addition, according to the present invention, it is possible to provide a video monitoring system, a method thereof, and an imaging apparatus that enable highly accurate incident detection in a short period of time.
In addition, according to the present invention, it is possible to provide a video monitoring system, a method thereof, and an imaging apparatus that make it possible to predict the occurrence of an incident in advance.

FIG. 1 is a configuration diagram of a video surveillance system according to an embodiment of the present invention. FIG. 2 is a diagram illustrating an example of the monitoring target site system illustrated in FIG. 1. It is a functional block diagram of the process which an incident process program provides. FIG. 4 is a functional block diagram of processing provided by the incident processing program. FIG. 5 is a diagram for explaining an application example of the video monitoring system shown in FIG. FIG. 6 is a diagram for explaining an operation example of the video surveillance system shown in FIG. FIG. 7 is a diagram for explaining a case where the video monitoring system shown in FIG. 5 is applied to building management. FIG. 8 is a diagram for explaining a process of predicting an incident in the video monitoring system shown in FIG. FIG. 9 is a diagram for explaining a video surveillance system according to a modification of the present invention.

Hereinafter, embodiments of the present invention will be described.
[overall structure]
FIG. 1 is a diagram illustrating an example of a video monitoring system 1 according to an embodiment of the present invention.
The video monitoring system 1 illustrated in FIG. 1 includes, for example, a monitoring target site system 100, an indent video management server 200, an expert terminal device 400, and an incident person terminal device 500, which communicate via a network 60.

  The video monitoring system 1 detects an incident based on the video data of the camera 140 of the monitoring target site system 100 installed at the monitoring target site. The incident video data and its attribute data stored in the indent video management server 200 are used for this incident detection.

[Monitored site system 100]
FIG. 2 is a diagram illustrating an example of the monitoring target site system 100 illustrated in FIG. 1.
As shown in FIG. 2, the monitoring target field system 100 includes, for example, a processing device 110, a camera 140, a sensor 150, and an administrator terminal device 160, which are connected to the network 60 via the hub 50.

In the monitoring target site system 100, a plurality of cameras 140 and sensors 150 are installed at a predetermined site.
The compressed video data corresponding to the image captured by the camera 140 and the sensor data from the sensor 150 are transmitted from the camera 140 to the processing device 110.

The sensor 150 is, for example, an optical sensor, a sound sensor, a vibration sensor, a temperature sensor, or the like, and is provided in the imaging range of the camera 140 or in the vicinity thereof.
The sensor 150 includes not only a physical sensor but also a logical sensor that analyzes and detects video data by the camera or the processing unit 113. In the present invention, the sensor 150 may not be used.

The processing device 110 includes a video recording unit 111, a processing unit 113, and a memory 115.
For example, the video recording unit 111 stores video data and sensor data received from the camera 140.
The processing unit 113 performs various processes based on the incident processing program 120 stored in the memory 115.
The memory 115 stores the incident processing program 120 and various data necessary for the processing of the processing unit 113.

FIG. 3 is a functional block diagram of processing provided by the incident processing program 120.
As illustrated in FIG. 3, the incident processing program 120 includes, for example, an incident attribute management unit 121, an incident detection unit 123, a notification unit 125, a level determination unit 127, and a pre-response processing unit 129.

  The processing unit 113 executes the incident processing program 120 and executes each function shown in FIG. 3 as follows.

The incident attribute management unit 121 generates incident video data corresponding to the incident detected by the incident detection unit 123.
Further, the incident attribute management unit 121 performs processing for associating the incident video data with the attribute data of the incident.
The incident attribute management unit 121 generates the incident attribute data based on the operation data from the administrator terminal device 160 shown in FIG.

At this time, the incident attribute data indicates an incident occurrence possibility level. The level may be input according to the operation of the administrator of the administrator terminal device 160, or may be calculated by a predetermined algorithm based on the input information.
In addition, the incident attribute management unit 121 generates the attribute data indicating the level as a numerical value based on conditions such as time, place, event holding, and congestion level related to the detected incident.

  The incident detection unit 123 uses video data captured by the camera 140 (video data read from the video recording unit 111) based on the incident video data and attribute data stored in the indent video management server 200 shown in FIG. Detect incidents.

  An incident is, for example, any interruption, obstruction, loss, emergency, danger, situation that can or will cause them.

The notification unit 125 notifies the occurrence of an incident to the contact address associated with the incident detected by the incident detection unit 123.
The notification unit 125 determines in advance a person in charge to perform notification and information on notification contents in accordance with the detected incident.
The association between the incident and the person in charge is determined in advance based on, for example, the attendance status, location, authority, skill, etc. of the person in charge.

The level determination unit 127 determines whether or not the possibility that an incident has occurred exceeds a certain level based on the incident video data and attribute data stored in the indent video management server 200.
Based on the incident video data and attribute data stored in the indent video management server 200, the level determination unit 127 determines whether or not the possibility that an incident will occur at a predetermined future timing exceeds a certain level.

When the level determination unit 127 determines that the possibility that an incident will occur exceeds a certain level, the prior response processing unit 129 performs processing associated with the incident in advance. Examples of the pre-response processing include door opening / closing, lighting on / off, temperature adjustment, alarm processing, and the like.
When the level determination unit 127 determines that the possibility that an incident will occur at a predetermined timing in the future exceeds a certain level, the pre-response processing unit 129 performs pre-processing associated with the incident in advance.

The manager terminal device 160 shown in FIG. 2 is used by a site manager related to the monitored site system 100.
The administrator terminal device 160 includes, for example, an operation unit 161, a display 163, and a processing unit 165.

  The site manager operates the operation unit 161 of the administrator terminal device 160 to display the video data of the camera 140 or the video data read from the video recording unit 111 of the processing device 110 on the display 163 for viewing, Performs operations for determining incident occurrence information by sound and determining attribute data related to the incident. The processing unit 165 transmits operation data corresponding to the operation of the operation unit 161 to the processing device 110.

[Indent video management server 200]
As illustrated in FIG. 1, the indent video management server 200 includes, for example, a video recording unit 211, a processing unit 213, and a memory 215.
The video recording unit 211 stores, for example, incident video data of incidents received from the monitored site system 100, attribute data, sensor data, and the like associated therewith.
As shown in FIG. 5, when there is a monitoring target site system 100 for each of a plurality of monitoring target sites, the video recording unit 211 includes incident video data from the plurality of monitoring target site systems 100 and attributes associated therewith. Data, sensor data, etc. are stored.

The processing unit 213 performs various processes based on the incident processing program 221 and the learning program 230 stored in the memory 215.
The memory 215 stores various data necessary for the incident processing program 221, the learning program 230, and the processing of the processing unit 113.

FIG. 4 is a functional block diagram of processing provided by the incident processing program 221.
As illustrated in FIG. 4, the incident processing program 221 includes, for example, an incident attribute management unit 221, an incident detection unit 223, a notification unit 225, a level determination unit 227, and a prior response processing unit 229.

The incident attribute management unit 221 records incident video data received from the monitoring target field system and attribute data associated therewith in the video recording unit 211.
Further, the incident attribute management unit 221 updates the attribute data recorded in the video recording unit 211 based on the operation data from the expert terminal device 400. As a result, it is possible to obtain highly accurate attribute data utilizing the experience of the expert.
At this time, the incident attribute data indicates an incident occurrence possibility level.
In addition, the incident attribute management unit 221 generates the attribute data based on conditions such as time, place, event holding, and congestion level related to the detected incident.

  The attribute data update by the expert terminal device 400 may be performed only when the level indicated by the attribute data associated with the operation of the administrator terminal device 160 exceeds a predetermined value, for example. This level indicates the possibility of an incident.

  The incident detection unit 223 detects an incident based on the incident video data and attribute data stored in the video recording unit 211. The incident is detected from a comprehensive viewpoint based on the incidents detected by the plurality of monitoring target field systems 100.

  The notification unit 225 notifies the occurrence of an incident to the contact address associated with the incident detected by the incident detection unit 223. For example, the notification unit 225 notifies the incident person in charge terminal device 500.

Based on the incident video data and attribute data stored in the video recording unit 211, the level determination unit 227 determines whether or not the possibility that an incident will occur exceeds a certain level.
Based on the incident video data and attribute data accumulated in the video recording unit 211, the level determination unit 227 determines whether or not the possibility that an incident will occur at a predetermined future timing exceeds a certain level.
The level determination unit 227 performs level determination not only on the site monitored by the specific monitoring target site system 100 but also on a comprehensively wide range.
In addition, in addition to the incident video data and attribute data, external information such as time, monitoring location, news, weather, and the like may be used for the level determination of the possibility that a future incident will occur by the level determination unit 227.

When the level determination unit 227 determines that the possibility that an incident will occur exceeds a certain level, the prior response processing unit 229 performs processing associated with the incident in advance.
When the level determination unit 227 determines that the possibility that an incident will occur at a predetermined timing in the future exceeds a certain level, the pre-response processing unit 229 performs pre-processing associated with the incident in advance.

The learning program 230 illustrated in FIG. 1 is executed by the processing unit 213 and realizes the following functions.
The learning program 230 generates learning data indicating the relationship between the incident and the attribute data based on the incident detected by the monitoring target site system 100 and the attribute data stored in the video recording unit 211.
The incident detection unit 123 shown in FIG. 3 of the monitoring target field system 100 further detects the incident using the learning data generated by the learning program 230.

Further, the incident attribute management units 121 and 221 specify the incident possibility level based on the incident video data after the pre-reaction processing units 129 and 229 execute the pre-process.
Further, the incident attribute management units 121 and 221 evaluate the preliminary processing executed by the preliminary response processing units 129 and 229 based on the identified incident occurrence possibility level and update the learning data. The prior correspondence processing units 129 and 229 further perform the preliminary processing using the updated learning data.

  The expert terminal device 400 and the incident person terminal device 500 are, for example, a smartphone, a mobile phone, a tablet terminal, or the like. Further, it may be a desktop or a portable computer.

[Overall processing]
Less than. An operation example of the video monitoring system 1 will be described.
FIG. 6 is a diagram showing an overview of the overall processing of the video monitoring system 1 shown in FIG.
Hereinafter, each step of FIG. 6 will be described.
Step S101:
For example, the camera 140 captures a moving image of an imaging range in the facility. Further, the processing unit 143 of the camera 140 records the captured moving image in the memory 141.
In addition, sensor data from the sensors 150 around the camera 140 is stored in the memory 141.

Step S102:
The camera 140 transmits the video data and sensor data output from the memory 141 to the processing device 110, and the processing device 110 receives this.

Step S103:
The processing device 110 records the video data and sensor data received from the camera 140 in the video recording unit 111 shown in FIG.

Step S104:
The incident detection unit 123 shown in FIG. 3 of the processing device 110 shown in FIG. 2 uses video data captured by the camera 140 based on the incident video data and attribute data stored in the indent video management server 200 shown in FIG. An incident is detected from (video data read from the video recording unit 111).

Step S105:
The incident attribute management unit 121 illustrated in FIG. 3 generates incident video data corresponding to the incident detected by the incident detection unit 123 in step ST104.
Further, the incident attribute management unit 121 performs processing for associating the incident video data with the attribute data of the incident.

Step S106:
The processing device 110 uploads (transmits) the incident video data, the attribute data, and the sensor data associated in step S105 to the indent video management server 200.
The indent video management server 200 receives these.

Step S107:
The notification unit 125 notifies the occurrence of an incident to the contact associated with the incident detected by the incident detection unit 123 in step S104. For example, the incident person terminal device 500 shown in FIG. 1 is notified.

Step S108:
The indent video management server 200 shown in FIG. 1 records the incident video data, attribute data, and sensor data received in step S6 in the video recording unit 211.

Step S109:
The expert operates the expert terminal device 400 to display the attribute data recorded in the video recording unit 211 in step ST108 on the display 403, and operates and updates the operation unit 401 as necessary.

Step S110:
The level determination unit 227 of the indent video management server 200 determines whether or not the possibility that an incident will occur exceeds a certain level based on the incident video data and attribute data stored in the video recording unit 211.

Step S111:
The level determination unit 227 of the indent video management server 200 shown in FIG. 4 is based on the incident video data and attribute data stored in the video recording unit 211, and the possibility that an incident will occur at a predetermined timing in the future exceeds a certain level. It is determined whether or not.
The level determination unit 227 performs level determination not only on the site monitored by the specific monitoring target site system 100 but also on a comprehensively wide range.

Step S112:
When the level determination unit 227 determines that the possibility of occurrence of an incident has exceeded a certain level, the prior response processing unit 229 of the indent video management server 200 illustrated in FIG. 4 performs processing associated with the incident in advance.
When the level determination unit 227 determines that the possibility that an incident will occur at a predetermined timing in the future exceeds a certain level, the pre-response processing unit 229 performs pre-processing associated with the incident in advance.

  As described above, according to the video monitoring system 1, the incident processing program 120 of the monitoring target field system 100 associates the attribute data with the incident video data and accumulates them in the indent video management server 200, and uses them to detect the incident. I do. Therefore, the incident can be detected with high accuracy in a short period of time based on the data accumulated in the indent video management server 200.

  In addition, according to the video monitoring system 1, a conventional monitoring person can perform the same monitoring from outside the facility, which is limited to performing the video monitoring work only in the management facility (monitoring target place). By doing so, one manager can monitor one facility at the same time. When the number of cameras 140 to be monitored increases, there is a limit to the ability of the person in charge of monitoring. However, using the video analysis function by the indent video management server 200 using artificial intelligence (AI), a video that seems to be a problem ( By automatically sensing the incident video) and notifying the person in charge (the manager terminal device 160), it is possible to monitor the video from a large number of cameras 140 even with a small number of people.

  When using artificial intelligence, a technique called deep learning (machine learning) has achieved high performance, and its use is expanding. When these techniques are used, in order to obtain sufficient performance, it is necessary to prepare a large number of captured images in advance and to learn the artificial intelligence software. The large amount of video that is the target is called a learning set. A learning set mainly indicates a video that is selected by a person and indicates what video it is. That is, the task of tagging the video is performed.

In particular, since an image determined as an incident has a low occurrence frequency, it is difficult to have a large number of learning sets. According to the video monitoring system 1, when a plurality of monitoring target field systems 100 are used, incident video data for a learning set is sequentially accumulated in the video recording unit 211, and the performance of artificial intelligence is improved by performing sequential learning. Effectively improve. That is, according to the video surveillance system 1, by continuing to operate the system every day, the installed facilities and regions can further reduce the security personnel, and the safety is dramatically improved.
Furthermore, it is possible not only to detect incidents but also to predict the risk of incidents in advance and to respond to them before they occur, which is highly effective for preventive safety.

[Application example]
The video monitoring system 1 illustrated in FIG. 1 illustrates the case where one monitoring target location is used. However, in this application example, the monitoring target site systems 100 are arranged in a plurality of monitoring target locations.
FIG. 5 is a diagram for explaining an application example of the embodiment of the present invention.
As shown in FIG. 5, in the video monitoring system 1a of this application example, the monitoring target field systems 100_1 to 100_5 are arranged in five monitoring target locations, respectively.

  In the video monitoring system 1a, the indent video management server 200 accumulates incident video data, attribute data, and sensor data from the monitoring target field systems 100_1 to 100_5 of a plurality of monitoring target locations. Therefore, compared to the case where data is received from one monitoring target field system 100, the learning set can be increased, and the learning accuracy can be improved in a short period of time.

FIG. 7 is a diagram for explaining a case where the video monitoring system 1a shown in FIG. 5 is applied to building management.
As shown in FIG. 7, the video monitoring system 1 is used for monitoring a building from the first floor to the fourth floor.
The monitoring target site systems 100_1 to 100_5 have a camera 140 installed at each monitoring target location in the building.
In the monitoring target site systems 100_1 to 100_5, each camera 140 displays a live video corresponding to the imaging result on the display 163 of the administrator terminal device 160.

  The monitoring target field systems 100_1 to 100_5 transmit incident video data, attribute data, and sensor data to the indent video management server 200 and store them in the video recording unit 211.

Each monitored site system 100_1 to 100_5 can generate an incident at each site based on the data stored in the video recording unit 211 shown in FIG. 1 by the function of the level determination unit 127 shown in FIG. A level determination indicating sex is performed.
Then, for example, the current level (crisis level) is displayed together with the live video on the display of the administrator terminal device 160 shown in FIG. 2 of each of the monitoring target field systems 100_1 to 100_5.
The incident video data and level may be displayed and displayed collectively on one terminal of the video monitoring system 1a.

The level determination unit 127 determines whether or not the possibility that an incident has occurred exceeds a certain level based on the incident video data and attribute data stored in the indent video management server 200.
Based on the incident video data and attribute data stored in the indent video management server 200, the level determination unit 127 determines whether or not the possibility that an incident will occur at a predetermined future timing exceeds a certain level.

When the level determination unit 127 determines that the possibility that an incident will occur exceeds a certain level, the prior response processing unit 129 performs processing associated with the incident in advance.
When the level determination unit 127 determines that the possibility that an incident will occur at a predetermined timing in the future exceeds a certain level, the pre-response processing unit 129 performs pre-processing associated with the incident in advance.

FIG. 8 is a diagram showing the time transition of the result of the level determination for a predetermined incident by one indent video management server 200.
As illustrated in FIG. 8, the level determination unit 127 calculates a predicted incident ability level at a predetermined time interval (for example, 2 hours) for a predetermined incident.
Further, the level determination unit 127 calculates the actual incident possibility level at predetermined time intervals.

  When the predicted incident occurrence possibility level calculated by the level determination unit 127 exceeds a predetermined attention level A or crisis level B, the prior response processing unit 129 executes the preliminary response processing assigned thereto. Thereby, the actual incident possibility level can be lowered from the predicted incident possibility level.

[Modification]
FIG. 9 is a diagram for explaining a video surveillance system 1b according to a modification of the present invention.
As shown in FIG. 9, in the video monitoring system 1 b, all or part of the functions of the monitoring target field system 100 described above are incorporated in the AI camera 340.

  As illustrated in FIG. 9, the AI camera 340 includes, for example, an imaging unit 351, a frame memory 350, a processing unit 353, a communication unit 355, a video compression unit 357, and a recording medium 359.

The imaging unit 351 stores video data corresponding to the imaging result in the frame memory 350 before compression.
The processing unit 353 executes the indent processing program 370. The indent processing program 370 has, for example, the function of each unit shown in FIG.
The processing unit 353 reads the uncompressed video data from the frame memory 350 and performs incident detection based on information (learning set) accumulated in the video recording unit 211 of the indent video management server 200.
In the present embodiment, since the processing unit 353 detects an incident based on the video data before compression, it can perform highly accurate incident detection.
In addition, the processing unit 353 attaches attribute data to the incident video data of the detected incident by a predetermined algorithm.

The video compression unit 357 compresses the video data read from the frame memory 350 and records it on the recording medium 359.
The communication unit 355 transmits the incident video data and attribute data detected by the processing unit 353 to the indent video management server 200.

  According to the video monitoring system 1b, since the AI camera 340 has the incident detection function, the processing device 110 is not necessary.

The present invention is not limited to the embodiment described above.
That is, those skilled in the art may make various modifications, combinations, subcombinations, and alternatives regarding the components of the above-described embodiments within the technical scope of the present invention or an equivalent scope thereof. For example, the above-described embodiments are not necessarily limited to those having all the configurations described. A part of the configuration of one embodiment can be replaced with the configuration of another embodiment. Further, the configuration of another embodiment can be added to the configuration of a certain embodiment.

The functions of the incident attribute management units 121 and 221 described above may be provided to only one of the monitoring target site system 100 and the indent video management server 200.
The functions of the incident detection units 123 and 223 may be provided to only one of the monitoring target site system 100 and the indent video management server 200.
The functions of the notification units 125 and 225 may be provided to only one of the monitoring target field system 100 and the indent video management server 200.

The functions of the level determination units 127 and 227 may be provided to only one of the monitoring target field system 100 and the indent video management server 200.
The functions of the advance handling processing units 129 and 229 may be provided to only one of the monitoring target site system 100 and the indent video management server 200.

  The present invention can be applied to a video system that detects an incident by acquiring a moving image captured by a camera.

DESCRIPTION OF SYMBOLS 1,1a ... Image | video monitoring system 100 ... Monitoring object field system 110 ... Processing apparatus 120 ... Incident processing program 121 ... Incident attribute management part 123 ... Incident detection part 125 ... Notification part 127 ... Level determination part 129 ... Advance response processing part 140 ... Camera 150 ... Sensor 160 ... Administrator terminal device 200 ... Indent video management server 221 ... Incident processing program 221 ... Incident attribute management unit 223 ... Incident detection unit 225 ... Notification unit 227 ... Level determination unit 229 ... Advance response processing unit 230 ... Learning Program 400 ... expert terminal device 500 ... incident person in charge terminal device

Claims (5)

A recording device including a video recording unit that records video data of video captured by the camera;
Incident attribute management means for associating incident video data of an incident detected based on the video data with attribute data of the incident;
A server for storing the incident video data and the attribute data associated with the incident attribute management means for a plurality of incidents;
Level determination means for determining whether or not the possibility of occurrence of the incident exceeds a certain level based on the video data and the attribute data stored in the server;
Incident detection means for detecting the incident based on the incident video data and the attribute data stored in the server, and video data of the video captured by the camera;
When the level determination unit determines that the possibility of occurrence of the incident exceeds a certain level, a pre-response processing unit that performs pre-processing associated with the incident in advance,
Learning means for generating learning data indicating a relationship between the incident and the attribute data based on the incident detected by the incident detection means and the attribute data stored in the server used for detecting the incident And
The incident attribute management means associates the attribute data indicating an incident occurrence possibility level input by an administrator who viewed the video data with the incident video data, and the incident attribute management means The incident occurrence possibility level is identified based on the incident video data after the advance processing is executed, and the advance processing unit executed based on the identified incident occurrence possibility level Update the learning data by evaluating the process,
The pre-corresponding processing means further uses the learning data generated by the learning means to perform the pre-processing,
The video monitoring system, wherein the incident detection means detects the incident by further using the learning data generated by the learning means. A plurality of incident attribute management means for processing video data captured by the cameras at different monitoring target locations;
The video monitoring system according to claim 1, wherein the server accumulates the incident video data and the attribute data received from the plurality of incident attribute management means. Expert attribute management means for associating the attribute data updated in accordance with an expert operation based on the incident video data and the attribute data received from the incident attribute management means and storing the attribute data in the server in association with the incident video data. The video surveillance system according to claim 1 or 2. The incident attribute management means generates the attribute data indicating the incident occurrence level by a numerical value based on at least one condition of time, place, event holding, and congestion related to the incident. 4. The video surveillance system according to any one of 3. A recording step of recording video data of a video captured by the camera in a video recording unit;
Incident attribute management process for associating incident video data of an incident detected based on the video data with attribute data of the incident;
For a plurality of incidents, a server for storing the incident video data and the attribute data associated in the incident attribute management step ;
Based on the video data and the attribute data stored in the server, a level determination step for determining whether or not the possibility of occurrence of the incident exceeds a certain level;
An incident detection step of detecting the incident based on the incident video data and the attribute data stored in the server, and video data of a video captured by the camera;
If the level determination step determines that the possibility that the incident will occur exceeds a certain level, a pre-response processing step that performs pre-processing associated with the incident in advance,
A learning step of generating learning data indicating a relationship between the incident and the attribute data based on the incident detected in the incident detection step and the attribute data stored in the server used for the detection of the incident And
The incident attribute management step associates the attribute data indicating an incident occurrence possibility level input by an administrator who viewed the video data with the incident video data,
In the incident attribute management step, the incident occurrence possibility level is specified based on the incident video data after the pre-processing is executed in the pre-response processing step, and the incident occurrence possibility level is determined based on the specified incident occurrence possibility level. And updating the learning data by evaluating the preliminary processing executed in the preliminary processing step ,
The pre-corresponding processing step further uses the learning data generated in the learning step to perform the pre-processing,
The video monitoring method in which the incident detection step detects the incident by further using the learning data generated in the learning step .

Images