Campus monitoring method and system based on regional activity popularity analysis

Abstract

The invention relates to the field of campus monitoring, and discloses a campus monitoring method and system based on regional activity popularity analysis, and the method comprises the steps: determining a first personnel density and a first activity type of a first region in a first time period according to the monitoring data of the first region, obtaining a first historical personnel density and a first historical activity type of the first area in the first time period through the personnel activity heat map; when the first activity type does not conform to the first historical activity type and the first personnel density is greater than or equal to the first historical personnel density, acquiring a second personnel density of the first region in a second time period subsequent to the first time period, obtaining a second historical personnel density of the first area in a second time period through the personnel activity heat map; and when the second personnel density is greater than or equal to the preset second personnel density, sending abnormal activity prompt information to the first area. The campus monitoring effect is improved through regional activity popularity analysis.

Description

Technical Field

[0001] This application relates to the technical field of campus monitoring. More specifically, it relates to a campus monitoring method and system based on regional activity heat analysis. Background Art

[0002] Current campus intelligent monitoring methods have been widely applied in scenarios such as security management, teaching assistance, and resource scheduling. By deploying technologies such as high-resolution cameras, face recognition, and behavior analysis algorithms, the system can achieve functions such as real-time abnormal behavior warning, automatic student attendance, and intrusion detection in key areas. Some universities have introduced AI middle platforms to integrate multi-source data to optimize campus environment management. These technologies have significantly improved campus security efficiency and provided data support for teaching evaluation and resource allocation.

[0003] Current campus monitoring technologies often rely on simple behavior patterns for campus monitoring. Current campus monitoring technologies usually rely on preset rules or shallow algorithms, and there are significant limitations in the recognition of campus monitoring. They can only recognize stereotyped behaviors (such as stationary and fast moving) in specific scenarios, but cannot understand the dynamic associations and behavior motivations in complex situations. For example, the gathering behavior in the corridor may also be mislabeled as a security hazard due to the lack of environmental analysis. In addition, the simple pattern recognition of campus monitoring is difficult to support accurate management decisions. This "one-size-fits-all" recognition method not only reduces the practicality of the system, but may also cause teachers and students to resist the monitoring technology due to frequent false alarms. Summary of the Invention

[0004] The purpose of this application is to provide a campus monitoring method and system based on regional activity heat analysis, which solves the technical problem of poor effectiveness of relying on simple behavior pattern recognition in campus monitoring, and achieves the technical effect of improving the effectiveness of campus monitoring through regional activity heat analysis.

[0005] A campus monitoring method based on regional activity heat analysis provided by an embodiment of the present application, the method includes: based on campus area monitoring data, determining a personnel activity heat map corresponding to each area, where the personnel activity heat includes the historical personnel density and historical activity types in each area at different time periods; through an activity detection model, according to the monitoring data of the first area, determining the first personnel density and the first activity type of the first area in the first time period, and obtaining the first historical personnel density and the first historical activity type of the first area in the first time period through the personnel activity heat map; when the first activity type does not conform to the first historical activity type, and when the first personnel density is greater than or equal to the first historical personnel density, obtaining the second personnel density of the first area in the second time period subsequent to the first time period, and obtaining the second historical personnel density of the first area in the second time period through the personnel activity heat map; when the second personnel density is greater than or equal to the preset second personnel density, sending an abnormal activity prompt message to the first area.

[0006] In a possible implementation manner, the method further includes: determining the activity detection models of the first area in different time periods through the monitoring data of the first area in different time periods; through the activity detection model, according to the monitoring data of the first area, determining the first personnel density and the first activity type of the first area in the first time period, including: determining the first activity detection model corresponding to the first area in the first time period, and determining the second activity detection model corresponding to the second area around the first area in the first time period; through the second activity detection model, according to the second monitoring data of the second area around the first area in the first time period, determining the second personnel density and the second activity type of the second area in the first time period; when each second activity type conforms to the second historical activity type, and when the second personnel density is less than the second historical personnel density, through the first activity detection model in the first time period, according to the monitoring data of the first area in the first time period, determining the first personnel density and the first activity type of the first area in the first time period.

[0007] In another possible implementation, through an activity detection model, based on the monitoring data of the first area, the first personnel density and the first activity type in the first time period of the first area are determined, and it further includes: when the second activity type corresponding to each second area does not conform to the preset second activity type, or when the second personnel density corresponding to each second area is greater than or equal to the preset second personnel density, through a model selector, based on the second activity type and the second personnel density, among the activity detection models in different time periods corresponding to the first area, determine the first replacement activity detection model corresponding to the first activity detection model in the first time period of the first area; where the number of second areas is greater than or equal to the preset number of second areas; through the first replacement activity detection model, based on the monitoring data of the first area in the first time period, determine the first personnel density and the first activity type in the first time period of the first area.

[0008] In another possible implementation, the method further includes: determining the preset area correlation between the second area and the first area, obtaining the weather state adjustment factor corresponding to the weather state of the second area, and determining the product of the second personnel density, the preset area correlation, and the weather state adjustment factor to adjust the second personnel density.

[0009] In another possible implementation, the method further includes: when the second activity type corresponding to each second area does not conform to the preset second activity type, or when the second personnel density corresponding to each second area is greater than or equal to the preset second personnel density, and obtaining the review result of the historical abnormal activity prompt information corresponding to different activity types sent by the first area through the first activity detection model in the preset historical time period, determining the false alarm rate of the review result; obtaining the replacement review result of the historical abnormal activity prompt information corresponding to different activity types sent by the first area through the first replacement activity detection model in the preset historical time period, determining the false alarm rate of the replacement review result; when the false alarm rate of the review result is greater than or equal to the false alarm rate of the replacement review result, through the first replacement activity detection model, based on the monitoring data of the first area in the first time period, determine the first personnel density and the first activity type in the first time period of the first area; when the false alarm rate of the review result is less than the false alarm rate of the replacement review result, through the first activity detection model, based on the monitoring data of the first area in the first time period, determine the first personnel density and the first activity type in the first time period of the first area.

[0010] In another possible implementation, the method further includes: obtaining the activity type and personnel density of the key area corresponding to the key area. When the activity type of the key area does not conform to the preset key area activity type, or when the personnel density of the key area is greater than or equal to the preset key area personnel density, the replacement activity detection model of all areas in the first time period is determined by the model selector according to the activity type and personnel density of the key area; where the key areas include square areas, assembly areas, and school gate areas.

[0011] In another possible implementation, the method further includes: obtaining the campus path change time and campus path change information of the campus area, and determining multiple path change areas corresponding to the campus path change information; updating the activity detection model of the target activity type corresponding to the multiple path change areas according to the monitoring data of the multiple path change areas in the preset historical time period after the campus path change time.

[0012] In another possible implementation, updating the activity detection model of the target activity type corresponding to the multiple path change areas includes: obtaining multiple similar areas of the multiple path change areas, and updating the activity detection model of the target activity type corresponding to the multiple path change areas according to the monitoring data in the similar areas; where the multiple similar areas are areas with similar pedestrian flow characteristics and path change areas.

[0013] In another possible implementation, the method further includes: obtaining the review results of the historical abnormal activity prompt information corresponding to different activity types sent by the first area and the second area in the preset historical time period, and determining the false alarm rate of the abnormal activity prompt information corresponding to different activity types sent by the first area and the second area; when the target false alarm rate corresponding to the target activity type is greater than or equal to the preset target false alarm rate, updating the activity detection model of the target activity type corresponding to the first area and the second area in the preset historical time period according to the monitoring data of the first area and the second area in the preset historical time period.

[0014] The embodiment of the present application also provides a campus monitoring system based on regional activity heat analysis, including a unit for executing the method described in any one of the above.

[0015] The beneficial effects of the embodiment of the present application compared with the prior art are:

[0016] An embodiment of the present application provides a campus monitoring method based on regional activity heat analysis. The method includes: determining a personnel activity heat map corresponding to each area based on campus area monitoring data, where the personnel activity heat includes the historical personnel density and historical activity types in each area at different time periods; through an activity detection model, determining the first personnel density and the first activity type in the first area during the first time period according to the monitoring data of the first area, and obtaining the first historical personnel density and the first historical activity type in the first area during the first time period through the personnel activity heat map; when the first activity type does not conform to the first historical activity type and when the first personnel density is greater than or equal to the first historical personnel density, obtaining the second personnel density in the second time period subsequent to the first time period in the first area, and obtaining the second historical personnel density in the second area during the second time period through the personnel activity heat map; when the second personnel density is greater than or equal to the preset second personnel density, sending an abnormal activity prompt message to the first area. In the embodiment of the present application, when the first activity type obtained through the personnel activity heat map in the first area does not conform to the first historical activity type and the first personnel density is greater than or equal to the first historical personnel density, the second personnel density in the second time period subsequent to the first time period in the first area can be obtained, and the second historical personnel density in the second area during the second time period can be obtained through the personnel activity heat map. When the first area meets the abnormal activity condition during the second time period, an abnormal activity prompt message is sent, improving the accuracy of monitoring the abnormal activity prompt in the first area. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a flowchart of the first campus monitoring method based on regional activity heat analysis provided by the embodiment of the present application;

[0019] Figure 2 It is a flowchart of the second campus monitoring method based on regional activity heat analysis provided by the embodiment of the present application;

[0020] Figure 3 It is a flowchart of the third campus monitoring method based on regional activity heat analysis provided by the embodiment of the present application;

[0021] Figure 4 It is a flowchart of the fourth campus monitoring method based on regional activity heat analysis provided by the embodiment of the present application;

[0022] Figure 5 This is a schematic diagram of the logical structure of a campus monitoring system provided by an embodiment of the present application based on regional activity heat analysis. Detailed implementation manners

[0023] It should be understood that when used in the specification of the present application and the appended claims, the term "comprising" indicates the presence of the described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their combinations.

[0024] It should also be understood that the term "and / or" used in the specification of the present application and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0025] As used in the specification of the present application and the appended claims, the term "if" can be interpreted as "when", "once", "in response to determining" or "in response to detecting" according to the context. Similarly, the phrase "if determined" or "if [the described condition or event] is detected" can be interpreted as meaning "once determined", "in response to determining", "once [the described condition or event] is detected" or "in response to detecting [the described condition or event]" according to the context.

[0026] In addition, in the description of the specification of the present application and the appended claims, the terms "first", "second", "third", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0027] The reference to "one embodiment" or "some embodiments" or the like described in the specification of the present application means that a specific feature, structure or characteristic described in connection with the embodiment is included in one or more embodiments of the present application. Thus, the statements "in one embodiment", "in some embodiments", "in other some embodiments", "in still other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "comprising", "including", "having" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0028] The simple pattern recognition of current campus monitoring is difficult to support accurate management decisions and may trigger the resistance of teachers and students to monitoring technology due to frequent false alarms.

[0029] For the above reasons, the embodiments of the present application provide a campus monitoring method based on regional activity heat analysis. The method includes: based on campus area monitoring data, determining a personnel activity heat map corresponding to each area, where the personnel activity heat includes the historical personnel density and historical activity types in each area at different time periods; through an activity detection model, according to the monitoring data of the first area, determining the first personnel density and the first activity type in the first time period of the first area, and obtaining the first historical personnel density and the first historical activity type of the first area in the first time period through the personnel activity heat map; when the first activity type does not conform to the first historical activity type and when the first personnel density is greater than or equal to the first historical personnel density, obtaining the second personnel density in the second time period subsequent to the first time period of the first area, and obtaining the second historical personnel density of the first area in the second time period through the personnel activity heat map; when the second personnel density is greater than or equal to the preset second personnel density, sending an abnormal activity prompt message to the first area. In the embodiments of the present application, when the first activity type obtained through the personnel activity heat map of the first area does not conform to the first historical activity type and the first personnel density is greater than or equal to the first historical personnel density, the second personnel density in the second time period subsequent to the first time period of the first area can be obtained, and the second historical personnel density of the first area in the second time period can be obtained through the personnel activity heat map. When the first area meets the abnormal activity condition in the second time period, an abnormal activity prompt message is sent, improving the accuracy of monitoring the abnormal activity prompt of the first area.

[0030] In some scenarios, a campus monitoring method based on regional activity heat analysis in the embodiments of the present application can be applied to the monitoring of campus areas, capable of monitoring abnormal activities in areas such as assembly areas, playground areas, corridor areas, and school gates, improving the accuracy of abnormal activity monitoring.

[0031] The following specifically describes a campus monitoring method based on regional activity heat analysis provided by the embodiments of the present application with specific examples.

[0032] Figure 1 It is a schematic flowchart of the first campus monitoring method based on regional activity heat analysis provided by the embodiments of the present application. As Figure 1 shown, the above method includes S110 to S130, and the following specifically describes S110 to S130.

[0033] S110. Based on the campus area monitoring data, determine the personnel activity heat map corresponding to each area. The personnel activity heat includes the historical personnel density and historical activity types in each area during different time periods. Through the activity detection model, according to the monitoring data of the first area, determine the first personnel density and the first activity type in the first area during the first time period, and obtain the first historical personnel density and the first historical activity type in the first area during the first time period through the personnel activity heat map.

[0034] When conducting campus monitoring, first, the intelligent cameras distributed in key areas are used to collect campus area monitoring data in real time to construct a dynamic data set in the spatio-temporal dimension. The personnel activity heat map not only contains the distribution curves of historical personnel density at different times, but also marks typical activity types such as classroom teaching, physical exercise, assembly activities, etc., providing an intuitive data reference for campus management.

[0035] Exemplarily, the personnel activity heat map of each area can be generated through the density estimation algorithm and the spatio-temporal clustering model.

[0036] In the embodiments of the present application, when real-time analysis of a specific area is required, the system will call the activity detection model based on deep learning. According to the monitoring data of the first area, determine the first personnel density and the first activity type in the first area during the first time period, and obtain the first historical personnel density and the first historical activity type in the first area during the first time period through the personnel activity heat map.

[0037] Taking the first teaching building as an example, during the monitoring period from 8 am to 10 am, the system can accurately identify behavior patterns such as teacher-student interaction in the classroom and short stays in the corridor through this model, and calculate the real-time personnel density index. Through the comparative analysis with historical data in the same period, the management personnel can not only quickly master the activity characteristics of the current area, but also predict the occurrence probability of abnormal aggregation events.

[0038] Exemplarily, the activity detection model can adopt a two-stream convolutional neural network (2D + 3D hybrid structure) to extract features from the video frame sequence, and combine the long short-term memory network (LSTM) to capture the time dependence of behavior patterns.

[0039] S120. When the first activity type does not conform to the first historical activity type, and when the first personnel density is greater than or equal to the first historical personnel density, obtain the second personnel density in the second time period subsequent to the first time period in the first area, and obtain the second historical personnel density in the second area in the second time period through the personnel activity heat map.

[0040] In this monitoring method, when it is determined based on the analysis of the real-time video stream that the first activity type in the first area during the first time period is different from the first historical activity type in the same historical period, and the real-time personnel density Dreal ≥ Dh, the system is further configured to start the high-density area tracking algorithm through the edge computing node. During the consecutive second time period after the end of the first time period, video frame sequences of this area are continuously collected at an interval of Δt (Δt ≤ 5 minutes), and at the same time, the multi-target tracking model is activated to dynamically track the movement trajectories of personnel, generating a three-dimensional spatio-temporal heat map containing position coordinates, velocity vectors, and residence durations. Through the three-dimensional spatio-temporal heat map, the second historical personnel density in the first area during the second time period can be obtained.

[0041] Exemplarily, when determining the second historical personnel density during the second time period, an improved YOLOv8 object detection model can be called to perform multi-scale feature extraction on the monitoring screen, and the Kalman filtering algorithm is combined to achieve personnel counting calibration in a dynamic scenario; a high-precision real-time heat map is generated through the Density Regression Network (DensityNet).

[0042] S130. When the second personnel density is greater than or equal to the preset second personnel density, an abnormal activity prompt message is sent to the first area.

[0043] During monitoring, when the system detects that the real-time personnel density D2real in the second time period satisfies that the second personnel density is greater than or equal to the preset second personnel density (D2real ≥ D2), the third-level abnormal determination mechanism is triggered, and an abnormal activity prompt message is sent to the first area.

[0044] Exemplarily, when a certain class borrows a self-study room temporarily for a meeting (the activity type changes from "self-study" to "meeting"), it is possible that the first activity type does not conform to the first historical activity type, and when the first personnel density is greater than or equal to the first historical personnel density. If it is detected in the second time period that the personnel leave quickly (the density returns to normal), the system determines it as a reasonable temporary activity and does not send an abnormal activity prompt message to the first area.

[0045] Exemplarily, if it is detected that there is "item transfer" (abnormal activity type) in the self-study room at night and the density continues to rise, it is possible that the first activity type does not conform to the first historical activity type, and when the first personnel density is greater than or equal to the first historical personnel density, and when it is detected in the second time period that the second personnel density is greater than or equal to the preset second personnel density, theft may occur, and the system sends an abnormal activity prompt message to the first area to improve the accuracy of monitoring the abnormal activity prompt in the first area.

[0046] The beneficial effect of the above-mentioned implementation method is that when the first activity type of the first area obtained through the personnel activity heat map does not conform to the first historical activity type and the first personnel density is greater than or equal to the first historical personnel density, the second personnel density of the first area in a second time period subsequent to the first time period is obtained; when the second personnel density is greater than or equal to the preset second personnel density, abnormal activity prompt information is issued to the first area, thereby improving the accuracy of monitoring abnormal activity prompts in the first area.

[0047] In some implementations, the method further includes: determining an activity detection model of the first area in different time periods through monitoring data of the first area in different time periods.

[0048] In order to accurately detect activities in different time periods, the activity detection model of the first area in different time periods can be determined through the monitoring data of the first area in different time periods, thereby achieving further selection of the activity detection model and improving the detection accuracy of the activity detection model.

[0049] Exemplarily, the historical monitoring data of the first area can be divided into time periods (such as hours, class schedule periods), and independent activity detection models are trained for each time period. For example, the morning class period model corresponds to the normal flow density (such as 200 people) and activity type (such as "orderly entry and exit of the classroom") of the learning teaching building from 8:00 to 10:00; the lunch break period model corresponds to the high-density flow (such as 500 people / hour) and activity type (such as "queueing for meals") of the learning cafeteria from 12:00 to 13:00.

[0050] Exemplarily, a plurality of different activity detection models (eg, “normal class model”, “break traffic model”, “emergency evacuation model”) may be trained for the first area in different time periods (eg, during classes, during breaks, and at night).

[0051] Exemplarily, when determining the activity detection model of the first area in different time periods through the monitoring data of the first area in different time periods, the number of monitored people, activity types, and the similarity of activity types in the first area in different time periods can be determined first, and then the activity detection model of the first area in different time periods can be determined based on the number of monitored people, activity types, and the similarity of activity types in the first area in different time periods to improve the accuracy of activity detection in the first area in different time periods.

[0052] Exemplarily, when determining the activity detection model of the first area in different time periods through monitoring data of the first area in different time periods, the activity detection model of the first area in different time periods may be determined through a deep learning model.

[0053] In some implementations, in the above S110, through an activity detection model, based on the monitoring data of the first area, determine the first personnel density and the first activity type of the first area within the first time period, including S111 to S112. The following is a specific description of S111 to S112.

[0054] S111. Determine the first activity detection model corresponding to the first area within the first time period, and determine the second activity detection model corresponding to the second area around the first area within the first time period.

[0055] When performing activity detection, through spatio-temporal correlation modeling and regional linkage verification mechanism, the detection models of the first area and the second area around the first area can be dynamically associated within the same time period to achieve more accurate anomaly judgment. Specifically, the first activity detection model corresponding to the first area within the first time period can be determined, and the second activity detection model corresponding to the second area around the first area within the first time period can be determined to improve the activity detection effect for different areas.

[0056] Exemplarily, within the same time period (for example, 10:00 - 11:00 in the morning), the activity patterns of the first area (such as a teaching building) and the second area (such as the corridors and stairwells around the teaching building) usually have strong correlations, and the activity detection models for different areas are different.

[0057] S112. Through the second activity detection model, based on the second monitoring data of the second area around the first area within the first time period, determine the second personnel density and the second activity type of the second area within the first time period. When each second activity type conforms to the second historical activity type, and when the second personnel density is less than the second historical personnel density, through the first activity detection model within the first time period, based on the monitoring data of the first area within the first time period, determine the first personnel density and the first activity type of the first area within the first time period.

[0058] In the embodiment of this method, the detection result of the first area can be verified through the real-time state (including activity type and personnel density) of the second area, and false judgments caused by isolated events and sensor noise can be excluded.

[0059] When performing activity detection, through the second activity detection model, based on the second monitoring data of the second area around the first area within the first time period, determine the second personnel density and the second activity type of the second area within the first time period, and then the activity of the first area can be detected through the second personnel density and the second activity type within the second area.

[0060] When performing activity detection on the first area, when each second activity type meets the second historical activity type, and when the second personnel density is less than the second historical personnel density, it indicates that the activity status in the second area has reviewed the historical status. Then, the first activity detection model in the first time period can be used to determine the first personnel density and the first activity type in the first time period according to the monitoring data of the first area in the first time period, thereby realizing activity detection of the first area in the first time period.

[0061] For example, a high density (60 people / hour) is detected in the classroom (first area), but the density in the adjacent corridor (second area) is 0 (the density during historical breaks should be 20 people / hour), that is, the density in the corridor is abnormally low and the activity type does not match (for example, "stationary" instead of "passing"). The system determines that the classroom detection result may be a false alarm (for example, camera obstruction) rather than a real gathering event.

[0062] The beneficial effect of the above implementation method is that when performing activity detection, the detection model of the first area and the second area around the first area can be dynamically associated within the same time period through spatiotemporal correlation modeling and regional linkage verification mechanism, thereby achieving more accurate abnormality judgment.

[0063] The beneficial effect of the above-mentioned implementation method is that by forcing the first area and the second area to follow a logical relationship (such as crowd flow path, density change law) within the same time period, the activity detection model is run on the first area only when the second area is in normal status, thereby improving the accuracy of activity monitoring in the first area.

[0064] In some implementations, in the above S110, the first personnel density and the first activity type in the first area within the first time period are determined through the activity detection model according to the monitoring data of the first area, and also include S113 to S114, and S113 to S114 are specifically described below.

[0065] S113. When the second activity type corresponding to each second area does not conform to the preset second activity type, or when the second personnel density corresponding to each second area is greater than or equal to the preset second personnel density, a first replacement activity detection model corresponding to the first activity detection model of the first area in the first time period is determined by a model selector according to the second activity type and the second personnel density in the activity detection models in different time periods corresponding to the first area. The number of the second areas is greater than or equal to the preset number of the second areas.

[0066] In the embodiments of the present application, through a multi-region anomaly collaborative detection dynamic model replacement mechanism, when anomalies occur in multiple second regions around the first region, a replacement detection model adapted to the current scenario can be automatically selected to improve the accuracy of anomaly recognition. Among them, the preset activity types of the second regions (for example, the activity type of the corridor should be "passage") and the preset personnel density thresholds (for example, corridor density ≤ 200 people / hour) can be defined.

[0067] It should be noted that the number of second regions is greater than or equal to the preset number of second regions, and the preset number of second regions can be 3 to ensure the verification accuracy of the first region through the activity types of multiple second regions.

[0068] During the detection, when the second activity type corresponding to each second region does not conform to the preset second activity type, or when the second personnel density corresponding to each second region is greater than or equal to the preset second personnel density, it indicates the consistency of abnormal activities in multiple second regions. Furthermore, local noise (such as a single sensor failure) can be distinguished from real global events (such as a fire, a gathering). Then, through the model selector, according to the second activity type and the second personnel density, among the activity detection models in different time periods corresponding to the first region, the first replacement activity detection model corresponding to the first activity detection model of the first region in the first time period is determined. Furthermore, according to the abnormal characteristics (such as activity type, density) of the second region, a pre-trained adapted model is selected to avoid the failure of a fixed model in an emergency.

[0069] Exemplarily, when "evacuation" is detected in multiple corridors and the density exceeds the limit, the first activity detection model can be switched to the first replacement activity detection model (for example, a fire escape model) to optimize the behavior recognition under smoke interference.

[0070] Exemplarily, the model selector can be a deep learning model trained with the labeled second activity type and second personnel density data.

[0071] S113. Through the first replacement activity detection model, according to the monitoring data of the first region in the first time period, determine the first personnel density and the first activity type of the first region in the first time period.

[0072] After obtaining the first replacement activity detection model, through the first replacement activity detection model, according to the monitoring data of the first region in the first time period, determine the first personnel density and the first activity type of the first region in the first time period to achieve accurate detection of the first personnel density and the first activity type.

[0073] The beneficial effect of the above implementation is that through the dynamic model replacement mechanism, when abnormalities occur in multiple second regions around the first region, the original detection model is replaced with the first replacement active detection model adapted to the current abnormal scenario, thereby improving the accuracy of the detection results in the first region.

[0074] In some implementations, the above method further includes: determining the preset regional correlation between the second region and the first region, obtaining the weather state adjustment factor corresponding to the weather state of the second region, and determining the product of the second personnel density, the preset regional correlation, and the weather state adjustment factor to adjust the second personnel density.

[0075] To improve the detection accuracy of the first region, a preset regional correlation can be set between the first region and the second region to represent the intensity of the pedestrian flow association between the first region and the second region.

[0076] Exemplarily, the preset regional correlation can be 0 - 1, and the preset regional correlation can be obtained based on historical data statistics or spatial topological relationship calculations.

[0077] Exemplarily, due to the direct flow of the pedestrian flow and high correlation, the preset regional correlation between the teaching building (the first region) and the adjacent corridor (the second region) can be 0.9.

[0078] Exemplarily, there is no direct association between the pedestrian flow of the teaching building and the distant playground, and the preset regional correlation can be 0.2.

[0079] Traditional methods directly use the original density, which may misjudge the low density of the playground as abnormal on rainy days. To further improve the detection accuracy of the first region, the weather state (such as rain, snow, fog) may affect the personnel density. Furthermore, the personnel density can be adjusted through the weather state adjustment factor, and the weather state adjustment factor corresponding to the weather state of the second region can be obtained.

[0080] After obtaining the original density, regional correlation, and weather adjustment factor, the product of the second personnel density, the preset regional correlation, and the weather state adjustment factor can be determined through the calculation formula of the corrected density = original density × regional correlation × weather adjustment factor to adjust the second personnel density.

[0081] The beneficial effect of the above implementation is that by introducing the regional correlation weight and the dynamic weather state adjustment factor, the personnel density in the second region can be corrected in multiple dimensions to more realistically reflect the actual pedestrian flow state in a complex environment, improving the activity monitoring accuracy of the second region and the first region.

[0082] The beneficial effects of the above implementation method are also as follows. Through the dual correction of the regional relevance weight and the weather dynamic adjustment factor, environmental and spatial interferences are eliminated, the real pedestrian flow state is approximated, high-relevance regions can be focused on simultaneously, redundant calculations are reduced, and the weather changes can be responded to in real time, improving the system robustness.

[0083] Figure 2 FIG. 4 is a schematic flowchart of a second campus monitoring method based on regional activity heat analysis provided by an embodiment of the present application. As Figure 2 shown, the above method further includes S210 to S220, and the following is a specific description of S210 to S220.

[0084] S210. When the second activity type corresponding to each second region does not conform to the preset second activity type, or when the second personnel density corresponding to each second region is greater than or equal to the preset second personnel density, obtain the review result of the historical abnormal activity prompt information corresponding to different activity types sent by the first region through the first activity detection model within the preset historical time period, and determine the review result false alarm rate corresponding to the review result. Obtain the replacement review result of the historical abnormal activity prompt information corresponding to different activity types sent by the first region through the first replacement activity detection model within the preset historical time period, and determine the replacement review result false alarm rate corresponding to the replacement review result.

[0085] To further improve the detection effect, the optimal detection model of the first region can be dynamically selected through historical false alarm rate comparison, and the best detection model can be selected from the original model or the replacement model. When the activity type in the second region is abnormal, the detected activity type in the second region does not conform to the preset (for example, the corridor should be "passing" but "staying" is detected), or the density in the second region exceeds the limit, and the real-time personnel density ≥ the preset threshold (for example, the corridor density ≥ 300 people, and the preset threshold is 200 people), the proportion of the abnormal prompt information sent by the original model (the first activity detection model) in the first region within the preset historical time period that is confirmed as a false alarm by manual review can be counted, and the review result false alarm rate corresponding to the review result corresponding to the first activity detection model can be obtained.

[0086] Exemplarily, if the first activity detection model triggers 100 alarms in the past 30 days, and 15 of them are false alarms, the review result false alarm rate of the first activity detection model is 15%.

[0087] At the same time, the historical abnormal activity prompt information corresponding to different activity types can be detected through the first replacement activity detection model. Furthermore, the false alarm proportion through the replacement model (the first replacement activity detection model) within the same time period can be counted, and the replacement review result false alarm rate corresponding to the replacement review result can be determined.

[0088] Exemplarily, if the first replacement activity detection model triggers 80 alarms during the same period, with 6 false alarms, the false alarm rate corresponding to the replacement review result can be determined to be 7.5%.

[0089] S220. When the false alarm rate of the review result is greater than or equal to the false alarm rate of the replacement review result, through the first replacement activity detection model, based on the monitoring data of the first area in the first time period, determine the first personnel density and the first activity type of the first area in the first time period. When the false alarm rate of the review result is less than the false alarm rate of the replacement review result, through the first activity detection model, based on the monitoring data of the first area in the first time period, determine the first personnel density and the first activity type of the first area in the first time period.

[0090] When the false alarm rate of the review result is greater than or equal to the false alarm rate of the replacement review result, it indicates that the false alarm rate of the original model ≥ the false alarm rate of the replacement model, showing that the replacement model has performed better historically. The system switches to the replacement model for detection. Specifically, through the first replacement activity detection model, based on the monitoring data of the first area in the first time period, determine the first personnel density and the first activity type of the first area in the first time period.

[0091] Conversely, when the false alarm rate of the review result is less than the false alarm rate of the replacement review result, through the first activity detection model, based on the monitoring data of the first area in the first time period, determine the first personnel density and the first activity type of the first area in the first time period. Through data-driven decision-making, it is ensured that the model selection always favors the solution with a lower false alarm risk, improving the accuracy of campus monitoring.

[0092] The beneficial effect of the above implementation method is that when the second area triggers an abnormal condition (inconsistent activity type or density exceeding the limit), the review results of the first activity detection model and the first replacement activity detection model are respectively determined, and their false alarm rates are respectively calculated. If the false alarm rate of the original model ≥ the false alarm rate of the replacement model, it indicates that the replacement model has performed better historically. The system switches to the replacement model for detection. Conversely, the original model is retained, realizing data-driven decision-making to ensure that the model selection always favors the solution with a lower false alarm risk and improving the accuracy of campus monitoring.

[0093] In some implementation methods, the above method further includes: obtaining the key area activity type and the key area personnel density corresponding to the key area. When the key area activity type corresponding to the key area does not conform to the preset key area activity type, or when the key area personnel density corresponding to the key area is greater than or equal to the preset key area personnel density, through the model selector, based on the key area activity type and the key area personnel density, determine the replacement activity detection model of all areas in the first time period. Among them, the key areas include the square area, the assembly area, and the school gate area.

[0094] In campus monitoring, key areas can be identified. Abnormal conditions in key areas may indicate global risks, and the detection strategy of the entire area can be adjusted in a coordinated manner to respond quickly. Specifically, by giving priority to monitoring key areas on campus (such as squares, assembly areas, and school gates), the activity types and personnel densities of key areas corresponding to key areas can be obtained, and their activity types and personnel densities can be analyzed in real time, and the global detection strategy can be dynamically adjusted based on the preset safety threshold.

[0095] When the activity type in the key area is abnormal (for example, "people gathering" is detected at the school gate during non-opening hours) or the density exceeds the limit (for example, the real-time density of the square exceeds the assembly capacity), that is, when the key area activity type corresponding to the key area does not meet the preset key area activity type, or when the key area personnel density corresponding to the key area is greater than or equal to the preset key area personnel density, the model selector can match the adapted replacement activity detection model for all areas according to the key area activity type and the key area personnel density (for example, start the "high-density control model" or the "illegal intrusion identification model").

[0096] It should be noted that key areas are preset and may include square areas, assembly areas and school gate areas.

[0097] The beneficial effect of the above implementation method is that through the pre-awareness of risks in key areas, the system's active defense capability is significantly improved, and the safety of key areas can be prioritized. For example, when there is an abnormality at the school gate, the access control model is quickly activated to avoid the spread of hidden dangers; the global model dynamic switching mechanism reduces computational redundancy while ensuring accuracy (it only takes effect globally when the key area triggers replacement), reducing computing resource consumption.

[0098] The beneficial effect of the above implementation method is that it is specially optimized for high-frequency risk scenarios (such as trampling at gatherings, illegal intrusions), and the false alarm rate is lower than that of the general model, which is especially suitable for the real-time security needs of large-scale campus scenarios.

[0099] Figure 3 A flow chart of a third campus monitoring method based on regional activity heat analysis provided in an embodiment of the present application is shown as follows: Figure 3 As shown, the above method also includes S310 to S320, and S310 to S320 are described in detail below.

[0100] S310: Obtain the campus path change time and campus path change information of the campus area, and determine multiple path change areas corresponding to the campus path change information.

[0101] When the campus path in the campus area changes, the campus path change time and campus path change information can be obtained, and multiple path change areas corresponding to the campus path change information can be determined, so as to perform more accurate abnormal activity recognition on the multiple path change areas.

[0102] Exemplarily, the affected path change areas (such as closed roads, newly added channels) can be identified through the campus path change time and topological information, and the monitoring data (including personnel flow patterns, density distributions, activity types) within a preset historical time period (such as within 30 days) after the path change in these areas can be extracted. For the target activity types of the path change areas (such as "one-way traffic" for the detour path, "aggregation risk" for the new channel), the activity detection models for the target activity types corresponding to the multiple path change areas are further updated to improve the activity detection effect for the target activity types of the path change areas.

[0103] S320. Update the activity detection models for the target activity types corresponding to the multiple path change areas according to the monitoring data of the multiple path change areas within a preset historical time period after the campus path change time.

[0104] After obtaining the multiple path change areas, the activity detection models for the target activity types corresponding to the multiple path change areas can be updated according to the monitoring data of the multiple path change areas within a preset historical time period after the campus path change time, so as to optimize the activity detection models for the multiple path change areas and improve the monitoring effect after the path change.

[0105] Exemplarily, when updating the activity detection models for the target activity types corresponding to the multiple path change areas, the parameters of the activity detection models can be updated and the feature weights can be adjusted. The incremental learning method is adopted to retain historical knowledge while integrating new scenario features, so that it adapts to the physical space logic after the change and avoids the computing power overhead of full model retraining.

[0106] The beneficial effect of the above implementation method is that the detection accuracy is improved in the changed area based on the directional data collection and model update of the path change event. The model update mechanism driven by the path change event supports the synchronous update of the independent detection models for multiple path change areas under multiple path changes, ensuring the real-time consistency of the campus global monitoring system.

[0107] In some implementation methods, in the above S320, updating the activity detection models for the target activity types corresponding to the multiple path change areas includes: obtaining multiple similar areas of the multiple path change areas, and updating the activity detection models for the target activity types corresponding to the multiple path change areas according to the monitoring data within the similar areas. Among them, the multiple similar areas are areas with similar human flow characteristics and path change areas.

[0108] To quickly update the independent detection models for multiple path change regions, the independent detection models for multiple path change regions can be updated based on cross-region feature transfer learning. Multiple similar regions of multiple path change regions are obtained, and the monitoring data (including pedestrian flow density, movement trajectories, and temporal characteristics of activity types) of the similar regions in the historical period are extracted. Using a transfer learning framework (such as a domain adaptation network), the knowledge of the similar regions is transferred to the target activity detection model of the path change region. Through the cross-region knowledge reuse and collaborative training mechanism, the model update efficiency and generalization ability are significantly improved.

[0109] Exemplarily, multiple similar regions (such as regions with a daily average flow fluctuation curve matching degree > 90%) highly similar to the pedestrian flow characteristics of the path change region can be identified through clustering analysis or similarity measurement (such as the dynamic time warping algorithm).

[0110] Exemplarily, the model parameters can be updated through joint training (such as optimizing the feature alignment loss function) to make the activity detection model adapt to the spatial topology and pedestrian flow distribution characteristics after the path change.

[0111] The beneficial effects of the above implementation method are that the dependence on the long-term data collection of the path change region itself is reduced through data migration of similar regions, and the model convergence speed is improved; the feature alignment technology effectively alleviates the distribution offset problem caused by the path change, and the detection accuracy of the target activity type (such as "abnormal aggregation") is guaranteed.

[0112] The beneficial effects of the above implementation method are also that it can support multi-region parallel migration, and the associated models of multiple path change regions can be updated at one time to ensure the spatio-temporal consistency of the global monitoring strategy.

[0113] Figure 4 FIG. is a schematic flowchart of a fourth campus monitoring method based on regional activity heat analysis provided by an embodiment of the present application. As Figure 4 shown, the above method further includes S410 to S420, and the following is a specific description of S410 to S420.

[0114] S410. Obtain the review results of the historical abnormal activity prompt information corresponding to different activity types sent by the first region and the second region within a preset historical time period, and determine the false alarm rates of the abnormal activity prompt information corresponding to different activity types sent by the first region and the second region.

[0115] To further improve the accuracy in campus monitoring, it is possible to obtain the review results of historical abnormal activity prompt information corresponding to different activity types issued in the first area and the second area within a preset historical time period, and based on the cross-region false alarm rate joint feedback mechanism, by statistically calculating the proportion of false alarms confirmed by manual or system review in the abnormal prompt information triggered for different activity types in the first area and the second area within the preset historical time period (where the false alarm rate = number of false alarms / total number of alarms), determine the false alarm rates of the abnormal activity prompt information corresponding to different activity types issued in the first area and the second area, and achieve the positioning of the target false alarm rates corresponding to the target activity types (such as "assembly" and "staying").

[0116] Exemplarily, the second area can be the second area around the first area, and the functional types of the second area and the first area are the same. For example, both the first area and the second area can be similar places such as assembly halls.

[0117] S420. When the target false alarm rate corresponding to the target activity type is greater than or equal to the preset target false alarm rate, update the activity detection model corresponding to the target activity type in the first area and the second area within the preset historical time period according to the monitoring data of the first area and the second area within the preset historical time period.

[0118] After obtaining the target false alarm rate corresponding to the target activity type, when the target false alarm rate ≥ the preset threshold (such as 10%), the historical monitoring data (including video streams and sensor timing data) of the first area and the second area can be extracted, and the activity detection model corresponding to the target activity type in the first area and the second area within the preset historical time period can be updated. Specifically, transfer learning or incremental learning methods can be used to re-calibrate the parameters and optimize the feature weights of the detection model for the target activity type, and focus on correcting the feature extraction layer and the classification decision layer that cause high false alarms.

[0119] The beneficial effects of the above implementation method are as follows: By optimizing the closed-loop model driven by false alarms, the accurate detection ability of specific activity types is improved, and the robustness of the model is optimized specifically; the dynamic update mechanism only iterates the model for activity types with high false alarm rates, reducing the consumption of computing resources compared with the full model update.

[0120] The beneficial effects of the above implementation method are also as follows: Through the second area around the first area, cross-region knowledge transfer (sharing the low false alarm feature extraction network of the second area to the first area) is carried out to enhance the generalization ability of the model in similar scenarios and improve the accuracy of model recognition.

[0121] The embodiment of the present application also provides a campus monitoring system based on regional activity heat analysis, including units for executing the method described in any one of the above.

[0122] Figure 5 The figure is a schematic logical structure diagram of a campus monitoring system based on regional activity heat analysis provided by an embodiment of the present application. As Figure 5 shown, the system 1 of this embodiment includes a processing unit 11, a storage unit 12, and a transceiver unit 13. The processing unit 11 is used to process data, the storage unit 12 is used to store data, and the transceiver unit 13 is used to send and receive data. The processing unit 11, the storage unit 12, and the transceiver unit 13 cooperate with each other to implement the above method. The beneficial effects of the embodiments of the present application have been described in the above method and will not be repeated here.

[0123] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units, due to being based on the same concept as the method embodiment of the present application, for their specific functions and the technical effects brought, reference can be specifically made to the method embodiment part and will not be repeated here.

[0124] Those skilled in the art can clearly understand that for the convenience and brevity of description, only the above division of each functional unit and module is used for illustration. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction and do not limit the protection scope of the present application. The specific working process of the units and modules in the above system can refer to the corresponding process in the foregoing method embodiment and will not be repeated here.

[0125] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, to implement all or part of the processes in the above-mentioned embodiment methods of this application, a computer program can be used to instruct relevant hardware to complete. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable medium can at least include: any entity or device that can carry the computer program code to the photographing device / terminal device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk, or an optical disc, etc. In some jurisdictions, according to legislation and patent practice, the computer-readable medium cannot be an electrical carrier signal and a telecommunication signal.

[0126] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0127] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or by a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0128] In the embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of the devices or units can be in an electrical, mechanical or other form.

[0129] The unit described as a separation component may or may not be physically separated. The component shown as a unit may or may not be a physical unit, that is, it may be located in one place or may be distributed over multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0130] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A campus monitoring method based on regional activity heat analysis, characterized in that, The method includes: Based on the campus area monitoring data, determine the personnel activity heat map corresponding to each area, where the personnel activity heat includes the historical personnel density and historical activity types in each area at different time periods; through the activity detection model, according to the monitoring data of the first area, determine the first personnel density and the first activity type in the first time period of the first area, and obtain the first historical personnel density and the first historical activity type in the first time period of the first area through the personnel activity heat map; When the first activity type does not match the first historical activity type, and when the first personnel density is greater than or equal to the first historical personnel density, obtain the second personnel density in the second time period following the first time period of the first area, and obtain the second historical personnel density in the second time period of the first area through the personnel activity heat map; When the second personnel density is greater than or equal to the preset second personnel density, send an abnormal activity prompt message to the first area.

2. The method according to claim 1, characterized in that, The method further includes: Determine the activity detection models of the first area at different time periods through the monitoring data of the first area at different time periods; Through the activity detection model, according to the monitoring data of the first area, determining the first personnel density and the first activity type in the first time period of the first area includes: Determine the first activity detection model corresponding to the first area in the first time period, and determine the second activity detection model corresponding to the second area around the first area in the first time period; Through the second activity detection model, according to the second monitoring data of the second area around the first area in the first time period, determine the second personnel density and the second activity type in the first time period of the second area; when each second activity type matches the second historical activity type, and when the second personnel density is less than the second historical personnel density, through the first activity detection model in the first time period, according to the monitoring data of the first area in the first time period, determine the first personnel density and the first activity type in the first time period of the first area.

3. The method according to claim 2, wherein Through the activity detection model, according to the monitoring data of the first area, determining the first personnel density and the first activity type in the first time period of the first area further includes: When each second activity type corresponding to the second area does not match the preset second activity type, or when each second personnel density corresponding to the second area is greater than or equal to the preset second personnel density, through the model selector, according to the second activity type and the second personnel density, among the activity detection models of the first area at different time periods, determine the first replacement activity detection model corresponding to the first activity detection model of the first area in the first time period; where the number of second areas is greater than or equal to the preset number of second areas; Through the first replacement activity detection model, according to the monitoring data of the first area in the first time period, determine the first personnel density and the first activity type in the first time period of the first area.

4. The method according to claim 3, wherein The method further includes: Determine the preset regional relevance between the second region and the first region, obtain the weather condition adjustment factor corresponding to the weather condition of the second region, and determine the product of the second personnel density and the preset regional relevance and the weather condition adjustment factor to adjust the second personnel density.

5. The method according to claim 4, characterized in that, The method further includes: When the second activity type corresponding to each second region does not conform to the preset second activity type, or when the second personnel density corresponding to each second region is greater than or equal to the preset second personnel density, obtain the review result of the historical abnormal activity prompt information corresponding to different activity types sent by the first region through the first activity detection model within the preset historical time period, and determine the false alarm rate of the review result; obtain the replacement review result of the historical abnormal activity prompt information corresponding to different activity types sent by the first region through the first replacement activity detection model within the preset historical time period, and determine the false alarm rate of the replacement review result corresponding to the replacement review result; When the false alarm rate of the review result is greater than or equal to the false alarm rate of the replacement review result, determine the first personnel density and the first activity type of the first region within the first time period according to the monitoring data of the first region within the first time period through the first replacement activity detection model; when the false alarm rate of the review result is less than the false alarm rate of the replacement review result, determine the first personnel density and the first activity type of the first region within the first time period according to the monitoring data of the first region within the first time period through the first activity detection model.

6. The method according to claim 5, wherein The method further includes: Obtain the key region activity type and the key region personnel density corresponding to the key region. When the key region activity type corresponding to the key region does not conform to the preset key region activity type, or when the key region personnel density corresponding to the key region is greater than or equal to the preset key region personnel density, determine the replacement activity detection model of all regions within the first time period according to the key region activity type and the key region personnel density through the model selector; wherein, the key regions include the square region, the assembly region, and the school gate region.

7. The method according to claim 6, characterized in that, The method further includes: Obtain the campus path change time and the campus path change information of the campus area, and determine the multiple path change regions corresponding to the campus path change information; Update the activity detection model of the target activity type corresponding to the multiple path change regions according to the monitoring data of the multiple path change regions within the preset historical time period after the campus path change time.

8. The method according to claim 7, wherein Updating the activity detection model of the target activity type corresponding to the multiple path change regions includes: Obtain multiple similar regions of the multiple path change regions, and update the activity detection model of the target activity type corresponding to the multiple path change regions according to the monitoring data within the similar regions; wherein, the multiple similar regions are regions with similar pedestrian flow characteristics and the path change regions.

9. The method according to claim 8, wherein The method further includes: Obtain the review results of the historical abnormal activity prompt information corresponding to different activity types sent by the first region and the second region within the preset historical time period, and determine the false alarm rate of the abnormal activity prompt information corresponding to different activity types sent by the first region and the second region; When the target false alarm rate corresponding to the target activity type is greater than or equal to the preset target false alarm rate, the activity detection model of the target activity type corresponding to the first area and the second area within the preset historical time period is updated according to the monitoring data of the first area and the second area within the preset historical time period.

10. A campus monitoring system based on regional activity heat analysis, characterized in that, It includes a unit for executing the method according to any one of claims 1 to 9.

Images