Big data-based camera monitoring data management system and method

Abstract

The invention discloses a camera monitoring data management system and method based on big data, and relates to the technical field of monitoring video management, and the method comprises the steps: obtaining a video monitoring record of a certain region, marking and classifying target objects in the video monitoring record, obtaining a video monitoring record of a first target object in a certain region, and carrying out the marking and classification of the target objects in the video monitoring record; collecting the video monitoring records into a first image sequence according to the motion trail of the first target object, calculating the encountering frequency of the first target object and the second type of target object in the first image sequence to obtain a first risk coefficient of the first monitoring area, and obtaining a second risk coefficient of the second monitoring area according to the motion trail of the second target object; and calculating the encountering frequency of a second target object and the first type of target objects in the second image sequence to obtain a second risk coefficient of the first monitoring area, classifying historical events in the historical record according to the numerical values of the first type of risk coefficient and the second type of risk coefficient, and performing risk prompt on a certain unit monitoring area.

Description

Technical Field

[0001] The present invention relates to the technical field of surveillance video management, and specifically to a camera surveillance data management system and method based on big data. Background Art

[0002] In the existing technical solutions, action recognition technology is used to recognize actions in videos, which can be applied to recognize dangerous actions in industrial production or monitor actions in case of accidents. However, due to the lack of spatio-temporal correlation in the application of action recognition technology in the existing technology, it is difficult to give early warnings of possible dangerous behaviors. In actual scenarios, some dangerous actions are not the subjective will of the person, but are subconscious decisions after being affected by the outside world. Therefore, it is very difficult to give early warnings of dangerous actions. Summary of the Invention

[0003] The purpose of the present invention is to provide a camera surveillance data management system and method based on big data to solve the problems raised in the existing technology.

[0004] To achieve the above purpose, the present invention provides the following technical solution: A camera surveillance data management method based on big data: Step S100: Obtain the video surveillance records of a certain area, mark and classify the target objects in the video surveillance records, set a certain action of a certain type of target as the target action, and recognize the target action in the video surveillance in a certain area through an action recognition algorithm; Step S200: Divide a certain area into several unit surveillance areas, obtain the unit surveillance areas where the target event occurs, obtain the first target object corresponding to the target event, obtain the video surveillance records of the unit surveillance areas, mark another target recognized from the video surveillance records as the second target object, and mark the category corresponding to the second target object as the second type of target; Step S300: Obtain the video surveillance records of the first target object in a certain area, gather the video surveillance records into a first image sequence according to the movement trajectory of the first target object, calculate the frequency of encounter between the first target object and the second type of target object in the first image sequence, and obtain the first risk coefficient of the first surveillance area; Step S400: Obtain the video surveillance records of the second target object in a certain area, gather the video surveillance records into a second image sequence according to the movement trajectory of the second target object, calculate the frequency of encounter between the second target object and the first type of target object in the second image sequence, and obtain the second risk coefficient of the first surveillance area; Step S500: Obtain the historical records of several target actions that occurred in a certain unit monitoring area. Classify the historical events in the historical records according to the numerical magnitudes of the first type of risk coefficient and the second type of risk coefficient to obtain the risk event set of a certain unit monitoring area. When the number of historical events in the risk event set exceeds the threshold, give a risk prompt for a certain unit monitoring area.

[0005] Further, step S100 includes: Step S101: Set a certain area in the city as the target area, divide the target area into several unit monitoring areas, and set at least one monitoring camera for each unit monitoring area; Step S102: Obtain the historical monitoring images of all monitoring cameras in the target area, mark the moving targets in the historical monitoring images, and classify the marked targets. Among them, at least two types of moving targets are obtained from the historical monitoring images; Step S103: Obtain a certain type of target as the first type of target, set a certain action of the first type of target as the target action, and apply the target recognition algorithm to recognize the target action in the video captured by the monitoring camera.

[0006] Further, step S200 includes: Step S201: Take a certain target action recognized by the monitoring camera as the target event, take the target that appears the target action in the target event as the first target object, take the unit monitoring area where the target event occurs as the first monitoring area, and record the monitoring camera corresponding to the first monitoring area as the target camera; Step S202: Obtain the monitoring video of the target camera. Before the target event occurs, in the time period of T0, obtain the targets that have a relative motion relationship with the first target object, and record them as the second target objects; Step S203: Obtain the target category corresponding to the second target object, and record the target category as the second type of target.

[0007] Further, step S300 includes: Step S301: Obtain the monitoring video records of the first target object in the target area before the target action appears. Arrange the monitoring videos according to the movement trajectory of the first target object in the target area to obtain the first image sequence VM1, where VM1: d1, d2, d3,... dn, and d1, d2, d3,... dn respectively represent the video records of the first target object in the 1st, 2nd, 3rd,... nth unit monitoring area before entering the first monitoring area; Step S302: Identify the second type of targets in the first image sequence, obtain the time duration during which there is a relative motion relationship between the second type of target objects and the first target object in the first image sequence, and denote the time duration as t1; Step S303: Obtain the time duration of all surveillance videos in the first image sequence, denote it as t0, and calculate the first risk coefficient α of the first surveillance area, α = t1 / t0; Obtain the video record of the movement trajectory of the first target object, extract the video segments in which the first target object and the second type of target appear simultaneously in the video record of the first target object, and obtain the time duration of the video segments; In the first image sequence, the more second type of target objects the first target object encounters, it indicates that the possibility of the second type of target objects themselves affecting the first target object to perform a target action is smaller. Since in the first surveillance area, the first target object has already performed a target action, then there is a possibility that due to the influence of the first surveillance area, the first target object is more likely to perform a target action; Therefore, the greater the first risk coefficient, the higher the possibility of a target action occurring in the first surveillance area.

[0008] Furthermore, step S400 includes: Step S401: Obtain all surveillance video records of the second target object before it enters the first surveillance area, arrange the surveillance videos according to the movement trajectory of the second target object in the target area, and obtain the second image sequence VM2, where VM2: e1, e2, e3,... em, and e1, e2, e3,... and em respectively represent the video records of the first, second, third,... mth unit surveillance areas of the second target object before it enters the first surveillance area; Step S402: Identify the first type of targets in the second image sequence, regard the video records in which the second target object and the first type of target exist simultaneously as target video records, obtain the number of target video records in the second image sequence, and denote the number as m0; Step S403: Calculate the second risk coefficient β of the first surveillance area, β = m0 / m; Obtain the video record of the movement trajectory of the second target object, extract the video segments in which the second target object and the first type of target appear simultaneously in the video record of the second target object, and obtain the number of encounters between the second type of target object and the first type of target in the video segments. Among them, when there is a picture in which the second type of target object and the first type of target appear simultaneously in the video segment, the video segment is regarded as a target video record. By counting the number of target video records in the second image sequence, the number of encounters between the second target object and the first type of target is obtained, that is, the unit surveillance areas where encounter relationships are generated; In the second image sequence, the more first - type target objects the second target object encounters, it indicates that the possibility of the second - type target object itself affecting the first target object to perform the target action is relatively small. Since in the first monitoring area, the first target object has performed the target action, there is a situation that due to the influence of the first monitoring area, the first target object is more likely to perform the target action. Therefore, when the second risk coefficient is larger, the first monitoring area has a higher possibility of the occurrence of the target action.

[0009] Further, step S500 includes: Step S501: Obtain several target events from the historical monitoring records of the first monitoring area, record them in the target event set R, respectively obtain the first image sequence and the second image sequence of each target event, calculate the first risk coefficient of each first image sequence, and the second risk coefficient of each second image sequence. Step S502: Set the first judgment threshold k1, record the target events with the first risk coefficient greater than the first judgment threshold in the first event set P1, and record the target events with the second risk coefficient less than the first judgment threshold in the second event set P2; Set the second judgment threshold k2, record the target events with the second risk coefficient greater than the second judgment threshold in the third event set P3, and record the target events with the second risk coefficient less than the second judgment threshold in the fourth event set P4; Step S503: Calculate the first reference set Q1, Q1 = P1 ∩ P3, the second reference set Q2, Q2 = P2 ∩ P4, obtain the number of target events in Q1 and record it as h1, and obtain the number of target events in Q2 and record it as h2; Step S504: When h1 > h2, set the first monitoring area as the risk - warning area, and give risk warnings to the first - type target and the second - type target entering the first monitoring area. Through two - stage screening, evaluate the historical events of the first monitoring area, and through the mutual verification of the two event sequences, to improve the accuracy of the risk assessment of the first monitoring area.

[0010] To better implement the above - mentioned method, a camera monitoring data management system based on big data is also proposed. The system includes: A regional monitoring management module, a target recognition module, a first risk - coefficient calculation module, a second risk - coefficient calculation module, and a risk judgment module. Among them, the regional monitoring management module is used to manage the monitoring videos in the target area, the target recognition module is used to identify the target objects in the video monitoring of the target area, the first risk - coefficient calculation module is used to calculate the first risk coefficient, the second risk - coefficient calculation module is used to calculate the second risk coefficient, and the risk judgment module is used to judge whether there is a risk in the unit monitoring area. Furthermore, the area monitoring and management module includes: a camera management unit, a historical video management unit, and a target management unit. Among them, the camera management unit is used to manage the monitoring cameras corresponding to the monitoring areas of each unit, the historical video management unit is used to manage the historical monitoring videos collected by the monitoring cameras, and the target management unit is used to manage the target objects in the monitoring videos; Furthermore, the target recognition module includes: a target event management unit, a target object recognition unit, and a target association unit. The target event management unit is used to identify target actions and mark target events. The target object recognition unit is used to identify the first target object and the second target object. The target association unit is used to obtain the association relationship between the target object and the target classification; Furthermore, the first risk coefficient calculation module includes: a first image sequence management unit, a time management unit, and a first risk coefficient calculation unit. Among them, the first image sequence management unit is used to manage the first image sequence, the time management unit is used to manage the time information in the video record, and the first risk coefficient calculation unit is used to calculate the first risk coefficient of the first monitoring area; Furthermore, the second risk coefficient calculation module includes: a second image sequence management unit, a target video management unit, and a second risk coefficient calculation unit. Among them, the second image sequence management unit is used to manage the second image sequence, the target video management unit is used to manage the target videos in the second image sequence, and the second risk coefficient calculation unit is used to calculate the second risk coefficient of the first monitoring area; Furthermore, the risk judgment module includes: a historical event management unit, an event classification unit, a reference set management unit, and a risk prompt unit. Among them, the historical event management unit is used to obtain all the historical events of the first monitoring area. The event classification unit is used to classify the historical events according to the judgment threshold. The reference set management unit is used to manage the first reference set and the second reference set. The risk prompt unit is used to give a risk prompt for a certain unit monitoring area when the number of historical events exceeds the threshold.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: When a target event occurs, by collecting path or trajectory information, two data sequences containing spatio-temporal relationships are obtained. Through the analysis of the data sequences, a judgment basis for the target actions that affect the unit area is obtained. Collect the records generated by the target events in the historical records, and improve the accuracy of the judgment through mutual verification. Description of the Drawings

[0012] Figure 1 It is a schematic structural diagram of a camera monitoring data management system based on big data according to the present invention; Figure 2 It is a schematic flow diagram of a camera monitoring data management method based on big data according to the present invention. Detailed implementation manners

[0013] Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0014] Embodiment: As Figure 1 and Figure 2 shown, the present invention provides a technical solution, a method for managing camera monitoring data based on big data: Step S100: Obtain the video monitoring records of a certain area, mark and classify the target objects in the video monitoring records, set a certain action of a certain type of target as the target action, and identify the target action in the video monitoring in a certain area through an action recognition algorithm; Among them, step S100 includes: Step S101: Set a certain area in the city as the target area, divide the target area into several unit monitoring areas, and set at least one monitoring camera corresponding to each unit monitoring area; Step S102: Obtain the historical monitoring images of all monitoring cameras in the target area, mark the moving targets in the historical monitoring images, and classify the marked targets. Among them, at least two types of moving targets are obtained from the historical monitoring images; Step S103: Obtain a certain type of target as the first type of target, set a certain action of the first type of target as the target action, and apply a target recognition algorithm to identify the target action in the video captured by the monitoring camera; In the embodiment, for example, the target action is that when the first type of target is a human body, the first target object is a certain person, and the human body falling is set as the target action; When the first type of target is a vehicle, the vehicle rolling over, sudden braking or sudden acceleration is set as the target action.

[0015] Step S200: Divide a certain area into several unit monitoring areas, obtain the unit monitoring area where the target event occurs, obtain the first target object corresponding to the target event, obtain the video monitoring records of the unit monitoring area, record another target identified from the video monitoring records as the second target object, and record the category corresponding to the second target object as the second type of target; Among them, step S200 includes: Step S201: Take a certain target action identified by the monitoring camera as the target event, take the target that appears in the target event as the first target object, take the unit monitoring area where the target event occurs as the first monitoring area, and record the monitoring camera corresponding to the first monitoring area as the target camera; Step S202: Obtain the surveillance video of the target camera. Before the target event occurs, within a time period of T0, obtain the target that has a relative motion relationship with the first target object, and denote it as the second target object; Step S203: Obtain the target category corresponding to the second target object, and denote the target category as the second type of target.

[0016] Step S300: Obtain the video surveillance record of the first target object in a certain area. According to the movement trajectory of the first target object, collect the video surveillance records into a first image sequence, calculate the frequency of encounter between the first target object and the second type of target object in the first image sequence, and obtain the first risk coefficient of the first surveillance area; Among them, step S300 includes: Step S301: Obtain the video surveillance record of the first target object in the target area before the target action appears. Arrange the surveillance videos according to the movement trajectory of the first target object in the target area to obtain a first image sequence VM1, where VM1: d1, d2, d3,... dn, and d1, d2, d3,... dn respectively represent the video records of the first target object in the 1st, 2nd, 3rd,... nth unit surveillance area before entering the first surveillance area; Step S302: Identify the second type of target in the first image sequence, obtain the time length during which the second type of target object has a relative motion relationship with the first target object in the first image sequence, and denote the time length as t1; Step S303: Obtain the time length of all surveillance videos in the first image sequence, denote it as t0, and calculate the first risk coefficient α of the first surveillance area, α = t1 / t0.

[0017] Step S400: Obtain the video surveillance record of the second target object in a certain area. According to the movement trajectory of the second target object, collect the video surveillance records into a second image sequence, calculate the frequency of encounter between the second target object and the first type of target object in the second image sequence, and obtain the second risk coefficient of the first surveillance area; Among them, step S400 includes: Step S401: Obtain all the video surveillance records of the second target object before entering the first surveillance area. Arrange the surveillance videos according to the movement trajectory of the second target object in the target area to obtain a second image sequence VM2, where VM2: e1, e2, e3,... em, and e1, e2, e3,... and em respectively represent the video records of the second target object in the 1st, 2nd, 3rd,... mth unit surveillance area before entering the first surveillance area; Step S402: Identify the first type of targets in the second image sequence. Use the video records in which both the second target object and the first type of targets exist in the video recording as target video records. Obtain the number of target video records in the second image sequence, and denote the number as m0. Step S403: Calculate the second risk coefficient β of the first monitoring area, where β = m0 / m.

[0018] Step S500: Obtain the historical records of several target actions that occurred in a certain unit monitoring area. Classify the historical events in the historical records according to the numerical values of the first type of risk coefficient and the second type of risk coefficient to obtain the risk event set of a certain unit monitoring area. When the number of historical events in the risk event set exceeds the threshold, give a risk warning for a certain unit monitoring area. Among them, Step S500 includes: Step S501: Obtain several target events from the historical monitoring records of the first monitoring area, include them in the target event set R. Respectively obtain the first image sequence and the second image sequence of each target event, and calculate the first risk coefficient of each first image sequence and the second risk coefficient of each second image sequence. Step S502: Set the first judgment threshold k1. Record the target events with the first risk coefficient greater than the first judgment threshold into the first event set P1, and record the target events with the second risk coefficient less than the first judgment threshold into the second event set P2. Set the second judgment threshold k2. Record the target events with the second risk coefficient greater than the second judgment threshold into the third event set P3, and record the target events with the second risk coefficient less than the second judgment threshold into the fourth event set P4. Step S503: Calculate the first reference set Q1, where Q1 = P1 ∩ P3, and the second reference set Q2, where Q2 = P2 ∩ P4. Obtain the number of target events in Q1 and denote it as h1, and obtain the number of target events in Q2 and denote it as h2. Step S504: When h1 > h2, set the first monitoring area as the risk warning area, and give risk warnings to the first type of targets and the second type of targets entering the first monitoring area.

[0019] The system includes: an area monitoring management module, a target recognition module, a first risk coefficient calculation module, a second risk coefficient calculation module, and a risk judgment module. Among them, the area monitoring management module is used to manage the monitoring videos in the target area. The area monitoring management module includes: a camera management unit, a historical video management unit, and a target management unit. The camera management unit is used to manage the monitoring cameras corresponding to each unit monitoring area, the historical video management unit is used to manage the historical monitoring videos collected by the monitoring cameras, and the target management unit is used to manage the target objects in the monitoring videos. Among them, the target recognition module is used to recognize target objects in the video monitoring of the target area. The target recognition module includes: a target event management unit, a target object recognition unit, and a target association unit. The target event management unit is used to recognize target actions and mark target events. The target object recognition unit is used to recognize the first target object and the second target object. The target association unit is used to obtain the association relationship between the target object and the target classification; Among them, the first risk coefficient calculation module is used to calculate the first risk coefficient. The first risk coefficient calculation module includes: a first image sequence management unit, a time management unit, and a first risk coefficient calculation unit. The first image sequence management unit is used to manage the first image sequence. The time management unit is used to manage the time information in the video record. The first risk coefficient calculation unit is used to calculate the first risk coefficient of the first monitoring area; Among them, the second risk coefficient calculation module is used to calculate the second risk coefficient. The second risk coefficient calculation module includes: a second image sequence management unit, a target video management unit, and a second risk coefficient calculation unit. The second image sequence management unit is used to manage the second image sequence. The target video management unit is used to manage the target video in the second image sequence. The second risk coefficient calculation unit is used to calculate the second risk coefficient of the first monitoring area; Among them, the risk judgment module is used to judge whether there is a risk in the unit monitoring area. The risk judgment module includes: a historical event management unit, an event classification unit, a reference set management unit, and a risk prompt unit. The historical event management unit is used to obtain all historical events in the first monitoring area. The event classification unit is used to classify historical events according to the judgment threshold. The reference set management unit is used to manage the first reference set and the second reference set. The risk prompt unit is used to give a risk prompt for a certain unit monitoring area when the number of historical events exceeds the threshold.

[0020] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for managing camera monitoring data based on big data, characterized in that: The method includes the following steps: Step S100: Obtain the video surveillance records of a certain area, mark and classify the target objects in the video surveillance records, set a certain action of a certain type of target as the target action, and identify the target action in the video surveillance in the said certain area through an action recognition algorithm; Step S200: Divide the said certain area into several unit surveillance areas, obtain the unit surveillance area where the target event occurs, obtain the first target object corresponding to the target event, obtain the video surveillance records of the unit surveillance area, mark another target identified from the video surveillance records as the second target object, and record the category corresponding to the second target object as the second type of target; Step S300: Obtain the video surveillance records of the first target object in the said certain area, collect the video surveillance records into a first image sequence according to the movement trajectory of the first target object, calculate the frequency of encounter between the first target object and the second type of target object in the first image sequence, and obtain the first risk coefficient of the first surveillance area; Step S400: Obtain the video surveillance records of the second target object in the said certain area, collect the video surveillance records into a second image sequence according to the movement trajectory of the second target object, calculate the frequency of encounter between the second target object and the first type of target object in the second image sequence, and obtain the second risk coefficient of the first surveillance area; Step S500: Obtain the historical records of several target actions that occurred in a certain unit surveillance area, classify the historical events in the historical records according to the numerical magnitudes of the first type of risk coefficient and the second type of risk coefficient, obtain the risk event set of the said certain unit surveillance area, and give a risk prompt for the said certain unit surveillance area when the number of historical events in the risk event set exceeds the threshold.

2. The method for managing camera monitoring data based on big data according to claim 1, wherein: Step S100 includes: Step S101: Set a certain area in the city as the target area, divide the target area into several unit surveillance areas, and set at least one surveillance camera corresponding to each unit surveillance area; Step S102: Obtain the historical surveillance images of all surveillance cameras in the target area, mark the moving targets in the historical surveillance images, and classify the marked targets, where at least two types of moving targets are obtained from the historical surveillance images; Step S103: Obtain a certain type of target as the first type of target, set a certain action of the first type of target as the target action, and apply a target recognition algorithm to identify the target action in the video captured by the surveillance camera.

3. A method for managing camera monitoring data based on big data according to claim 2, characterized in that: Step S200 includes: Step S201: Take a certain target action identified by the surveillance camera as the target event, take the target that appears in the target event as the first target object, take the unit surveillance area where the target event occurs as the first surveillance area, and record the surveillance camera corresponding to the first surveillance area as the target camera; Step S202: Obtain the surveillance video of the target camera, and in the time period of T0 before the occurrence of the target event, obtain the target that has a relative movement relationship with the first target object, and record it as the second target object; Step S203: Obtain the category of the target corresponding to the second target object, and denote the target category as the second type of target.

4. A method for managing camera monitoring data based on big data according to claim 3, characterized in that: Step S300 includes: Step S301: Obtain the surveillance video records of the first target object in the target area before the target action occurs. Arrange the surveillance videos according to the movement trajectory of the first target object in the target area to obtain the first image sequence VM1, where VM1: d1, d2, d3, ……, dn, and d1, d2, d3, ……, dn respectively represent the video records of the first target object in the 1st, 2nd, 3rd, ……, nth unit surveillance area before entering the first surveillance area; Step S302: Identify the second type of target in the first image sequence, and obtain the time length during which the second target object has a relative motion relationship with the first target object in the first image sequence, and denote the time length as t1; Step S303: Obtain the time length of all surveillance videos in the first image sequence, denoted as t0, and calculate the first risk coefficient α of the first surveillance area, α = t1 / t0.

5. A method for managing camera monitoring data based on big data according to claim 4, characterized in that: Step S400 includes: Step S401: Obtain all the surveillance video records of the second target object before entering the first surveillance area. Arrange the surveillance videos according to the movement trajectory of the second target object in the target area to obtain the second image sequence VM2, where VM2: e1, e2, e3, ……, em, and e1, e2, e3, ……, and em respectively represent the video records of the second target object in the 1st, 2nd, 3rd, ……, mth unit surveillance area before entering the first surveillance area; Step S402: Identify the first type of target in the second image sequence, regard the video records in which both the second target object and the first type of target exist as the target video records, obtain the number of target video records in the second image sequence, and denote the number as m0; Step S403: Calculate the second risk coefficient β of the first surveillance area, β = m0 / m.

6. The method for managing camera monitoring data based on big data according to claim 5, characterized in that: Step S500 includes: Step S501: Obtain several target events from the historical surveillance records of the first surveillance area, include them in the target event set R, respectively obtain the first image sequence and the second image sequence of each target event, and calculate the first risk coefficient of each first image sequence and the second risk coefficient of each second image sequence; Step S502: Set the first judgment threshold k1, record the target events with the first risk coefficient greater than the first judgment threshold into the first event set P1, and record the target events with the second risk coefficient less than the first judgment threshold into the second event set P2; Set the second judgment threshold k2, record the target events with the second risk coefficient greater than the second judgment threshold into the third event set P3, and record the target events with the second risk coefficient less than the second judgment threshold into the fourth event set P4; Step S503: Calculate the first reference set Q1, Q1 = P1 ∩ P3, the second reference set Q2, Q2 = P2 ∩ P4, obtain the number of target events in Q1 and denote it as h1, and obtain the number of target events in Q2 and denote it as h2; Step S504: When h1 > h2, set the first monitoring area as the risk warning area, and give risk warnings to the first type of targets and the second type of targets entering the first monitoring area.

7. A camera monitoring data management system based on big data, which is used to execute the method for managing camera monitoring data based on big data according to any one of claims 1-6, and is characterized in that: The system includes: a regional monitoring and management module, a target recognition module, a first risk coefficient calculation module, a second risk coefficient calculation module, and a risk judgment module. Among them, the regional monitoring and management module is used to manage the monitoring videos in the target area, the target recognition module is used to identify the target objects in the video monitoring of the target area, the first risk coefficient calculation module is used to calculate the first risk coefficient, the second risk coefficient calculation module is used to calculate the second risk coefficient, and the risk judgment module is used to judge whether there is a risk in the unit monitoring area.

8. The data management system for camera monitoring based on big data according to claim 7, characterized in that: The regional monitoring and management module includes: a camera management unit, a historical video management unit, and a target management unit. Among them, the camera management unit is used to manage the monitoring cameras corresponding to each unit monitoring area, the historical video management unit is used to manage the historical monitoring videos collected by the monitoring cameras, and the target management unit is used to manage the target objects in the monitoring videos; The target recognition module includes: a target event management unit, a target object recognition unit, and a target association unit. The target event management unit is used to identify target actions and mark target events. The target object recognition unit is used to identify the first target object and the second target object. The target association unit is used to obtain the association relationship between the target object and the target classification.

9. The data management system for camera monitoring based on big data according to claim 7, characterized in that: The first risk coefficient calculation module includes: a first image sequence management unit, a time management unit, and a first risk coefficient calculation unit. Among them, the first image sequence management unit is used to manage the first image sequence, the time management unit is used to manage the time information in the video record, and the first risk coefficient calculation unit is used to calculate the first risk coefficient of the first monitoring area; The second risk coefficient calculation module includes: a second image sequence management unit, a target video management unit, and a second risk coefficient calculation unit. Among them, the second image sequence management unit is used to manage the second image sequence, the target video management unit is used to manage the target videos in the second image sequence, and the second risk coefficient calculation unit is used to calculate the second risk coefficient of the first monitoring area.

10. A camera monitoring data management system based on big data according to claim 7, characterized in that: The risk judgment module includes: a historical event management unit, an event classification unit, a reference set management unit, and a risk warning unit. Among them, the historical event management unit is used to obtain all historical events in the first monitoring area, the event classification unit is used to classify historical events according to the judgment threshold, the reference set management unit is used to manage the first reference set and the second reference set, and the risk warning unit is used to give a risk warning to a certain unit monitoring area when the number of historical events exceeds the threshold.

Images