Artificial Intelligence-Based Campus Abnormal Behavior Detection System

Abstract

The present invention discloses a campus abnormal behavior detection system based on artificial intelligence, which relates to the technical field of abnormal behavior detection. The system includes: connecting to the campus monitoring platform to determine the first monitoring blind area and the first monitoring neighborhood; establishing a first standard aggregation feature set for the first monitoring blind area; obtaining the first monitoring video data; identifying and labeling personnel to generate the first monitoring object time series; performing aggregation feature recognition to generate the first real-time aggregation feature; performing abnormal aggregation feature recognition to generate the first abnormal probability; and when the first abnormal probability is greater than the preset warning probability, generating the first abnormal warning information. The present invention solves the technical problem in the prior art that traditional campus monitoring is difficult to cover blind areas and cannot accurately judge abnormal aggregation behaviors at different times, and achieves the technical effect of realizing the effective detection and warning of abnormal behaviors in the campus monitoring blind area and improving the campus safety management level.

Description

Technical Field

[0001] The present invention relates to the technical field of abnormal behavior detection, and particularly to a campus abnormal behavior detection system based on artificial intelligence. Background Art

[0002] In today's society, campus security issues have always been the focus of high attention from all parties. With the continuous development of education and the gradual expansion of campus scale, the campus environment has become increasingly complex. Although the traditional campus monitoring system has ensured campus security to a certain extent, there are many limitations. On the one hand, there are inevitably some monitoring blind spots on campus. Due to special locations or equipment layouts, etc., these areas are difficult to be directly covered by conventional monitoring devices. In these monitoring blind spots, once abnormal behaviors occur, such as student conflicts, illegal intrusions, etc., the traditional monitoring system often cannot detect them in time, which brings great potential safety hazards to campus security. On the other hand, the traditional monitoring system lacks effective analysis means for the personnel activity situations in different time zones on campus. At different time periods, the personnel gathering situations on campus vary greatly. For example, during the break between classes, the personnel are relatively concentrated, while during class time, they are relatively dispersed. If the personnel gathering characteristics cannot be accurately analyzed, it is difficult to judge whether there are abnormal gathering behaviors and it is impossible to take corresponding measures for intervention in time.

[0003] The prior art has the technical problems that traditional campus monitoring is difficult to cover blind spots and cannot accurately judge abnormal gathering behaviors at different times. Summary of the Invention

[0004] This application provides a campus abnormal behavior detection system based on artificial intelligence, which is used to solve the technical problems in the prior art that traditional campus monitoring is difficult to cover blind spots and cannot accurately judge abnormal gathering behaviors at different times.

[0005] In view of the above problems, this application provides a campus abnormal behavior detection system based on artificial intelligence.

[0006] This application provides a campus abnormal behavior detection system based on artificial intelligence, and the system includes:

[0007] The first monitoring blind area determination module is used to connect to the campus monitoring platform, determine the first monitoring blind area, and the first monitoring neighborhood connected to the first monitoring blind area; the first standard aggregation feature set establishment module is used to establish the first standard aggregation feature set of the first monitoring blind area, where the first standard aggregation feature set includes aggregation features of multiple time zones; the first monitoring video data acquisition module is used to acquire the first monitoring video data monitored by the monitoring devices in the first monitoring neighborhood; the first monitoring object time sequence generation module is used to identify and label the personnel entering the first monitoring blind area from the first monitoring neighborhood based on the first monitoring video data, and generate the first monitoring object time sequence; the first real-time aggregation feature generation module is used to perform aggregation feature recognition based on the first monitoring object time sequence and generate the first real-time aggregation feature; the first abnormal probability generation module is used to compare the first real-time aggregation feature with the first standard aggregation feature set, perform abnormal aggregation feature recognition, and generate the first abnormal probability; the first abnormal warning information generation module is used to generate the first abnormal warning information when the first abnormal probability is greater than the preset warning probability.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] The first monitoring blind area determination module is used to connect to the campus monitoring platform, determine the first monitoring blind area, and the first monitoring neighborhood connected to the first monitoring blind area; the first standard aggregation feature set establishment module is used to establish the first standard aggregation feature set of the first monitoring blind area; the first monitoring video data acquisition module is used to acquire the first monitoring video data monitored by the monitoring devices in the first monitoring neighborhood; the first monitoring object time sequence generation module is used to identify and label personnel and generate the first monitoring object time sequence; the first real-time aggregation feature generation module is used to perform aggregation feature recognition based on the first monitoring object time sequence and generate the first real-time aggregation feature; the first abnormal probability generation module is used to compare the first real-time aggregation feature with the first standard aggregation feature set, perform abnormal aggregation feature recognition, and generate the first abnormal probability; the first abnormal warning information generation module is used to generate the first abnormal warning information when the first abnormal probability is greater than the preset warning probability. It achieves the technical effect of effectively detecting and warning abnormal behaviors in the campus monitoring blind area and improving the campus security management level. Description of the Drawings

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0011] Figure 1 FIG. is a schematic structural diagram of a campus abnormal behavior detection system based on artificial intelligence provided by an embodiment of the present application;

[0012] Figure 2 FIG. is a schematic flowchart of determining the first monitoring blind area of a campus abnormal behavior detection system based on artificial intelligence provided by an embodiment of the present application.

[0013] Explanation of reference numerals: The first monitoring blind area determination module 10, the first standard aggregation feature set establishment module 20, the first monitoring video data acquisition module 30, the first monitoring object time series generation module 40, the first real-time aggregation feature generation module 50, the first abnormal probability generation module 60, and the first abnormal warning information generation module 70. Detailed implementation manners

[0014] The present application provides a campus abnormal behavior detection system based on artificial intelligence, which is used to solve the technical problems in the prior art that traditional campus monitoring is difficult to cover blind areas and cannot accurately judge abnormal aggregation behaviors at different times.

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application.

[0016] Embodiment, as Figure 1 shown, the present application provides a campus abnormal behavior detection system based on artificial intelligence, and the system includes:

[0017] The first monitoring blind area determination module 10, and the first monitoring blind area determination module 10 is used to connect to the campus monitoring platform, determine the first monitoring blind area, and the first monitoring neighborhood connected to the first monitoring blind area.

[0018] Specifically, the first monitoring blind area determination module 10 is responsible for connecting to the campus monitoring platform. By establishing a connection with the campus monitoring platform, the module can obtain the layout information, monitoring range, real-time monitoring data, and other relevant system parameters of each monitoring device on campus. This connection is achieved through network communication technology to ensure that the module can receive information from the monitoring platform in a timely and accurate manner. Connecting to the campus monitoring platform provides a data basis for determining the first monitoring blind area and the first monitoring neighborhood. Determining the first monitoring blind area is one of the core tasks of the module. There may be some areas on campus that cannot be directly covered by monitoring devices for various reasons, and these areas are defined as monitoring blind areas. The module will comprehensively consider factors such as the installation location, monitoring angle, coverage range, and building layout of the monitoring devices, and determine the specific location and scope of the first monitoring blind area through algorithm analysis or manual judgment. For example, analyze the field of view of the monitoring devices to find areas that are not covered by any monitoring device; or consider the occlusion situation of the buildings to determine possible blind area locations. Determining the first monitoring blind area is crucial for campus security management because these areas may become hidden points for potential safety hazards and require special attention. At the same time, the module also needs to determine the first monitoring neighborhood connected to the first monitoring blind area. Taking the first monitoring blind area as the center, analyze which surrounding monitoring areas are adjacent to the blind area, and determine the neighborhood by calculating the distance relationship between the coverage range of the monitoring device and the blind area. For example, if there is a certain overlap between the coverage range of a monitoring device and the boundary of the blind area, then the area corresponding to this monitoring device can be considered as part of the first monitoring neighborhood. The purpose of determining the first monitoring neighborhood is to use the monitoring data of the neighborhood to infer the situation of the blind area. Since the blind area cannot be directly monitored, by analyzing information such as personnel activities and environmental changes in the neighborhood, the possible situations in the blind area can be indirectly understood, thereby improving the accuracy and comprehensiveness of abnormal behavior detection.

[0019] The first standard aggregation feature set establishment module 20 is used to establish the first standard aggregation feature set of the first monitoring blind area, where the first standard aggregation feature set includes aggregation features of multiple time zones.

[0020] Specifically, the first standard aggregation feature set establishment module 20 is dedicated to constructing a first standard aggregation feature set for the first monitoring blind area. This feature set covers the aggregation features of multiple time zones, specifically including the aggregated number of people under normal circumstances and the characteristics of personnel changes over time. To obtain these features, the module collects the regular behavior samples of the first monitoring blind area in multiple time zones. These samples are from the historical data of the campus monitoring platform, including monitoring videos and personnel activity records at different time periods. By analyzing these samples, the normal behavior patterns of the first monitoring blind area at different times are understood. Then, based on these regular behavior samples, feature extraction is carried out. For determining the threshold of the aggregated number of people, the module counts the range of the number of people appearing in the samples of different time zones. For example, in the time zone before morning classes, a certain number of students pass through the first monitoring blind area on their way to the classroom. By analyzing multiple samples before morning classes, the threshold of the aggregated number of people in this time zone can be determined. This threshold represents the maximum number of people that may gather in the first monitoring blind area of this time zone under normal circumstances. For the stable feature of the change rate of the aggregated number of people, it represents the stability of the change rate of the aggregated number of people. The module calculates the change rate of the number of people over time in the samples of different time zones. If in a certain time zone, the change in the number of people is relatively gentle and the change rate fluctuates little, then it can be considered that the stable feature of the change rate of the aggregated number of people in this time zone is high. On the contrary, if the change in the number of people is drastic and the change rate fluctuates greatly, then the stable feature is low. By analyzing multiple samples, the stable feature of the change rate of the aggregated number of people in each time zone can be determined.

[0021] The first monitoring video data acquisition module 30 is used to acquire the first monitoring video data monitored by the monitoring devices within the first monitoring neighborhood.

[0022] Specifically, the main function of the first monitoring video data acquisition module 30 is to acquire the first monitoring video data monitored by the monitoring devices within the first monitoring neighborhood. Since the first monitoring blind area cannot be directly monitored, acquiring the monitoring video data of its neighborhood can provide an important information source for inferring the situation of the blind area. First, the module will establish connections with the monitoring devices within the first monitoring neighborhood through network communication protocols or specific monitoring system interfaces. Once the connections are successfully established, the module can receive the video data from these monitoring devices in real time. Next, the module will perform preliminary processing and storage on the acquired video data, such as compressing and encoding the video to reduce the data volume and facilitate subsequent analysis and processing. At the same time, the video data will be stored in a specific database or storage device for ready access and analysis. By acquiring the monitoring video data of the first monitoring neighborhood, on the one hand, the video of the neighborhood provides information about personnel flow. If there are people entering the first monitoring blind area from the neighborhood, their movement trajectories can be traced by analyzing the neighborhood video, thereby speculating on their possible behaviors in the blind area. On the other hand, the neighborhood video can also reflect changes in the environment. For example, if there are abnormal crowd gatherings, noises, or other abnormal phenomena in the neighborhood, it means that there may be potential problems in the first monitoring blind area. The first monitoring video data acquisition module 30 provides important basic data for the campus abnormal behavior detection system by acquiring the first monitoring video data monitored by the monitoring devices within the first monitoring neighborhood, which helps to improve the accuracy and timeliness of detecting abnormal behaviors in the first monitoring blind area.

[0023] The first monitoring object time series generation module 40, which identifies and labels the people entering the first monitoring blind area from the first monitoring neighborhood based on the first monitoring video data, and generates the first monitoring object time series.

[0024] Specifically, based on the first monitored video data, the first monitored object time series generation module 40 focuses on identifying and labeling the personnel entering the first monitored blind area from the first monitored neighborhood, and then generates the first monitored object time series. First, it receives the first monitored video data, which comes from the monitoring devices in the first monitored neighborhood and records the dynamics of the neighborhood and its surrounding areas. Based on the neural network model of deep learning, it accurately detects the positions and contours of the personnel in the video frame. Once the personnel are identified, it further determines whether these personnel enter the first monitored blind area from the first monitored neighborhood, and makes a comprehensive judgment by combining information such as the moving directions of the personnel in the video, position changes, and the layout of the monitoring devices. For the personnel determined to enter the first monitored blind area from the first monitored neighborhood, they are labeled, and the label includes the identity information of the personnel (if identifiable), entry time, entry position, etc. Through the labeling, the specific situation of each personnel entering the blind area can be clearly recorded. Finally, based on the labeled information, the first monitored object time series is generated. This time series is a record arranged in chronological order, which details the situations of the personnel entering the first monitored blind area from the first monitored neighborhood at different time points. For example, the time series can record that person A enters the blind area from a certain position in the neighborhood at time point T1, and person B enters the blind area at time point T2, etc. By analyzing this time series, the rules, frequencies of personnel entering the blind area, and the relationships between personnel can be understood. If the number of personnel entering the blind area increases abnormally or a specific combination of personnel appears during a certain period, it means that there may be potential abnormal situations. At the same time, the time series can also help trace the occurrence process of abnormal behaviors and provide strong evidence for subsequent investigations and handling.

[0025] The first real-time aggregation feature generation module 50, and the first real-time aggregation feature generation module 50 performs aggregation feature recognition based on the first monitored object time series to generate the first real-time aggregation feature.

[0026] Specifically, first, the generated first monitored object time series is received. This time series records the situations of the personnel entering the first monitored blind area from the first monitored neighborhood at different time points, providing the basic data for the recognition of aggregation features. Then, the recognition of aggregation features begins. The aggregation features mainly reflect the personnel aggregation situation in the first monitored blind area at a specific moment. For the generation of the first real-time aggregation features, multiple aspects are considered. On the one hand, the number of personnel in the first monitored blind area at the current moment is counted. By analyzing the first monitored object time series, it is determined how many personnel are in the blind area at a specific time point. This is an important aggregation feature index. On the other hand, the distribution and dynamic changes of the personnel are concerned. In addition, combining the time factor, the residence time of the personnel in the blind area and the change trend of the number of personnel over time are analyzed. If a large number of personnel pour into the blind area and stay for a long time in a short period, it means that an abnormal situation has occurred. Finally, through the comprehensive analysis of these factors, the first real-time aggregation features are generated. These features are represented by a feature vector or a set of descriptive indicators, such as the number of personnel at the current moment, the personnel distribution density, the personnel change rate, etc. These real-time aggregation features will be compared with the first standard aggregation feature set to determine whether there is an abnormal situation in the first monitored blind area.

[0027] The first abnormal probability generation module 60 is used to compare the first real-time aggregation features with the first standard aggregation feature set, identify abnormal aggregation features, and generate the first abnormal probability.

[0028] Specifically, obtain the first real-time aggregation feature, which is an immediate reflection of the current personnel aggregation situation in the first monitored blind area, including information such as the current number of people, distribution status, and change trend. At the same time, also obtain the already established first standard aggregation feature set, which covers the normal aggregation features of the first monitored blind area in multiple time zones and provides a standard reference for judging whether the current situation is abnormal. Then, compare the first real-time aggregation feature with the features in the corresponding time zone of the first standard aggregation feature set. Compare the current actual number of people with the normal aggregation personnel number threshold in the standard set for this time zone. If it exceeds the threshold range, there is an abnormality. Also, compare the personnel change rate with the stable feature of the personnel change rate under normal circumstances. If the change rate shows a large fluctuation or deviates from the normal range, there is an abnormality. Through this comparison, identify the abnormal aggregation features. If there is a large difference between the real-time feature and the standard feature, it indicates that an abnormal situation has occurred in the current first monitored blind area. These abnormal features are manifested as an abnormal increase or decrease in the number of people, an abnormal concentration or dispersion of the personnel distribution, and an abnormal fluctuation in the personnel change rate. Finally, generate the first abnormal probability according to the identification result of the abnormal aggregation feature. This probability value reflects the likelihood of an abnormal situation occurring in the first monitored blind area. When the abnormal probability is high, send a warning message in a timely manner to remind relevant personnel to pay close attention to the first monitored blind area so as to take corresponding measures to ensure campus safety.

[0029] The first abnormal warning message generation module 70 is used to generate a first abnormal warning message when the first abnormal probability is greater than a preset warning probability.

[0030] Specifically, if the first abnormal probability exceeds the preset warning probability, it indicates that a relatively serious abnormal situation may have occurred in the first monitored blind area. At this time, this module will be quickly activated to generate a first abnormal warning message. This warning message is presented in various forms, such as popping up a warning window on the display screen of the campus monitoring center, sending text messages or emails to notify relevant management personnel, or triggering an alarm sound, etc. The content of the warning message includes the location where the abnormality occurred (i.e., the first monitored blind area), the type of abnormality (judged according to the specific situation of the abnormal probability, such as excessive personnel aggregation, abnormal personnel change rate, etc.), and the recommended measures to be taken. In this way, after receiving the warning message, relevant personnel can quickly understand the specific situation of the abnormal situation and take corresponding countermeasures in a timely manner, such as dispatching security personnel to the scene to check, activating the emergency plan, which helps to promptly discover and handle abnormal situations on campus and ensure the safety of teachers and students and the normal order of the campus.

[0031] In a possible implementation manner, as Figure 2 shown, the first monitored blind area determination module 10 further includes:

[0032] Centering around the first monitoring blind spot, identify the boundary monitoring devices located within the first monitoring blind spot; obtain the monitoring parameters of the boundary monitoring devices; determine the boundary monitoring range and monitoring direction based on the monitoring parameters, and generate the first monitoring neighborhood.

[0033] Specifically, in the campus abnormal behavior detection system, centering around the first monitoring blind spot, boundary detection needs to be carried out first. This process is based on the campus monitoring platform. By analyzing the field of view of existing monitoring cameras, the exact boundary of the blind spot is determined. Next, identify the monitoring devices located at the edge of the blind spot, including cameras and sensors, etc., and record their positions and viewing angles. By integrating the information of these devices, compare the field of view of the boundary monitoring devices with the first monitoring blind spot, so as to identify the monitoring areas that can cover the blind spot. Finally, this identification result will be fed back to the monitoring system to optimize the monitoring layout, ensure effective coverage of the monitoring blind spot, and thus improve the campus security prevention ability.

[0034] After identifying the boundary monitoring devices, obtain the monitoring parameters of these devices, such as the viewing angle, focal length, resolution, and height of the camera, etc. These parameters will be used to judge whether the monitored object enters or exits the first monitoring blind spot, analyze the viewing range of the device, and combine the movement trajectory of the object in the video to determine its relative position. If the moving direction of the object is consistent with the monitoring range of the boundary device, its behavior can be determined as entering or leaving the blind spot. In this way, the system can monitor the campus security status in real time and issue early warnings in a timely manner.

[0035] Based on the obtained monitoring parameters of the boundary monitoring devices, analyze the viewing angle and monitoring range of each device to determine its coverage area. By calculating the monitoring directions of each device, identify the overlapping monitoring areas to form a complete first monitoring neighborhood. On this basis, combined with the height and focal length of the device, further optimize the boundary of the neighborhood to ensure no blind spots. In addition, the generated first monitoring neighborhood will be marked as a monitorable area, which can effectively track the objects entering or leaving the blind spot, thereby improving the accuracy and response ability of the overall security monitoring.

[0036] In a possible implementation manner, the first standard aggregation feature set establishment module 20 further includes:

[0037] Obtain the regular behavior samples of the first monitoring blind spot in the multiple time zones; perform feature extraction based on the regular behavior samples to generate the aggregation features of the multiple time zones, where the aggregation features include the aggregation personnel quantity threshold and the stable feature of the aggregation personnel change rate; the stable feature of the aggregation personnel change rate represents the stability of the change rate of the aggregated number of people.

[0038] Specifically, in order to obtain the regular behavior samples of the first monitoring blind area in multiple time zones, data mining is carried out by analyzing historical monitoring video data. First, the monitoring videos in different time periods are extracted, and the common activity types in the blind area are identified and classified, such as personnel passing by, gathering or loitering, etc. Then, the behavior frequency, duration and occurrence pattern of each time zone are recorded to establish a regular behavior database. Through these samples, a reference standard is provided for subsequent anomaly detection, so as to effectively identify the behavior patterns deviating from the normal, and further improve the response ability to abnormal behaviors.

[0039] Based on the obtained regular behavior samples, feature extraction is carried out to generate the aggregation features of multiple time zones. First, the behavior data in different time periods are analyzed to extract the number of aggregated people and its changing trend. For each time zone, the number threshold of aggregated people is calculated, that is, the minimum number of people regarded as aggregation in this time zone. This threshold is set according to the average level and standard deviation of historical data to ensure that it can accurately reflect the regular behavior. In addition, the change rate of aggregated people is analyzed to extract stable features. This includes calculating the increase and decrease rate of the number of aggregated people in a specific time period to identify normal fluctuations and abnormal sharp increases or decreases. Through these aggregation features, a dynamic monitoring model is established to effectively identify abnormal aggregation behaviors in real-time monitoring and provide timely warnings for campus safety.

[0040] The stable feature of the change rate of aggregated people is used to measure the smoothness and consistency of the change in the number of aggregated people. This feature reflects the regularity of the aggregation behavior by analyzing the change rate of the number of aggregated people in a specific time period. The number of aggregated people recorded at different time points (such as every minute, every hour) will be used to calculate the change rate, that is, the increase and decrease of the number of people per unit time. In order to quantify the stability, statistical methods are used to calculate the standard deviation of the change rate. If the standard deviation of the change rate is small, it means that the fluctuation of the number of aggregated people in this time zone is small and the change is relatively stable; on the contrary, if the standard deviation is large, it indicates that the change in the number of aggregated people is relatively drastic and there is a risk of abnormal behavior. In addition, a threshold of the change rate is established, and the change rate exceeding this threshold will be regarded as unstable, which indicates abnormal aggregation or personnel flow. By monitoring this stable feature, potential safety hazards can be better identified and strong data support can be provided for campus management.

[0041] In a possible implementation manner, the first abnormal probability generation module 60 further includes:

[0042] Obtain the first monitoring time zone of the first monitoring video data; match the target standard aggregation feature according to the first monitoring time zone in the first standard aggregation feature set; compare the first real-time aggregation feature with the matched target standard aggregation feature, establish an aggregation feature deviation, and combine with the monitored object to fuse the abnormal weight to generate the first abnormal probability.

[0043] Specifically, first extract relevant surveillance video data from the monitoring device. This process ensures that the acquired data is complete and unaltered. Analyze the timestamp information in the video data to determine the specific recording time period, which is achieved by extracting the metadata of the video file. Classify the timestamps according to the set time range, thereby dividing a specific first monitoring time zone. For example, divide a day into different time periods such as morning, afternoon, and evening. Finally, confirm the identified monitoring time zone to prepare for subsequent aggregated feature matching and abnormal behavior detection. This process ensures an accurate grasp of the monitoring time zone and helps improve the effectiveness of subsequent data analysis.

[0044] According to the first monitoring time zone, analyze the first standard aggregated feature set to match the target standard aggregated feature. First, extract the real-time aggregated features in this time zone, including the current number of people gathered, activity types, and change trends, etc. Then, compare these real-time features with the historical behavior patterns in the first standard aggregated feature set to find the most similar features. This matching process involves calculating the similarity between features and using the Euclidean distance algorithm to determine the degree of difference between the current aggregated features and historical normal behaviors. Through this matching, identify the target standard aggregated feature as a reference benchmark for subsequent abnormal detection. This process not only helps to clarify the aggregated behaviors in the current monitoring time zone but also provides key data support for identifying potential abnormalities, thereby enhancing the intelligence and response capabilities of the monitoring system.

[0045] Determine the first real-time aggregation feature and the target standard aggregation feature that matches the current time zone. For the first monitoring blind area, clarify the real-time aggregation features such as the actual number of people and the personnel change rate at the current moment. At the same time, find the standard aggregated personnel number threshold and the standard personnel change rate stability feature corresponding to the time zone from the first set of standard aggregation features. Then, calculate the aggregation feature deviation. Calculate the personnel number deviation, which is the absolute value of the difference between the actual number of people and the standard personnel number threshold; and the personnel change rate deviation, which is the absolute value of the difference between the actual personnel change rate and the standard personnel change rate stability feature. Then, combine these two deviations in the way of Euclidean distance to obtain the aggregation feature deviation value. For each monitoring object entering the first monitoring blind area, determine the anomaly weight according to its historical campus abnormal behavior records. Statistically, the ratio of the number of times of abnormal behavior to the total number of behavior records in the history of each monitoring object is used as the anomaly weight of the monitoring object. After that, calculate the sum of the anomaly weights of all monitoring objects, and then divide the anomaly weight of each monitoring object by this sum for normalization processing. When calculating the first anomaly probability, set the importance coefficient of the aggregation feature deviation and the anomaly weight respectively, and the sum of the two coefficients is 1. For each monitoring object, set an adjustment factor determined according to factors such as the severity of historical abnormal behavior. The first anomaly probability is equal to the aggregation feature deviation multiplied by its importance coefficient, plus the product of the normalized anomaly weights of all monitoring objects and the adjustment factor, multiplied by the importance coefficient of the anomaly weight. In this way, use the weighted average algorithm to accurately calculate the probability of abnormal conditions in the first monitoring blind area, providing strong support for campus safety management.

[0046] In a possible implementation manner, the first anomaly probability generation module 60 further includes:

[0047] Initialize the anomaly weights for all objects entering the first monitoring blind area in the time series of the first monitoring object to generate an initialized anomaly weight set; identify the initial anomaly probability based on the aggregation feature deviation, where the initial anomaly probability is proportional to the magnitude of the aggregation feature deviation; and perform a weighted influence calculation on the initial anomaly probability with the initialized anomaly weight set to generate the first anomaly probability.

[0048] Specifically, initialize the anomaly weights for all objects entering the first monitoring blind area in the time series of the first monitoring object. At this stage, assign an initial anomaly weight to each object entering the blind area. This weight is determined based on some basic factors, such as the type of the object (student, teacher, staff, etc.), the frequency of entering the blind area, etc. By initializing the anomaly weights of all objects, an initialized anomaly weight set is generated.

[0049] Next, the initial abnormal probability is identified based on the aggregation feature deviation. The aggregation feature deviation is established by comparing the first real-time aggregation feature with the matching target standard aggregation feature, which reflects the difference between the current state and the normal state of the first monitoring blind area. The initial abnormal probability is proportional to the magnitude of the aggregation feature deviation, that is, the larger the deviation, the higher the initial abnormal probability. This is because a larger deviation means that the situation in the current blind area is quite different from the normal situation, and the possibility of an abnormality occurring increases accordingly.

[0050] Finally, the initial abnormal probability is calculated with the initialized abnormal weight set to generate the first abnormal probability. In this step, the abnormal weight of each object affects the initial abnormal probability. If an object has a high abnormal weight, its contribution to the first abnormal probability will be relatively large. Through this weighted calculation, the influence degree of different objects on the abnormal situation in the blind area can be more accurately reflected. For example, if some objects are often associated with abnormal situations in history, their high abnormal weights will increase the first abnormal probability accordingly, thus attracting more attention from the system. By initializing the abnormal weights of the objects entering the first monitoring blind area, identifying the initial abnormal probability based on the aggregation feature deviation, and then performing the weighted influence calculation, a more accurate and targeted first abnormal probability is generated, providing an important basis for campus abnormal behavior detection and early warning.

[0051] In a possible implementation manner, the first abnormal probability generation module 60 further includes:

[0052] Obtain a historical campus abnormal behavior library, where any abnormal behavior sample in the historical campus abnormal behavior library includes an abnormal personnel identification sample; based on the abnormal personnel identification sample, count the first abnormal behavior frequency feature of the first person who has had an abnormal behavior; configure the initialization of the abnormal weight for the first person based on the first abnormal behavior frequency feature to generate a first abnormal weight; when the object entering the first monitoring blind area includes the first person, add the first abnormal weight to the initialized abnormal weight set.

[0053] Specifically, obtaining the historical campus abnormal behavior library is an important link. This historical campus abnormal behavior library stores relevant information on various abnormal behaviors that have occurred in the past. Among them, any abnormal behavior sample includes an abnormal personnel identification sample, which clearly indicates the specific personnel participating in the abnormal behavior. By obtaining this historical campus abnormal behavior library, analyzing and studying past abnormal behaviors, it is possible to better identify and predict possible future abnormal situations. For example, by analyzing the abnormal personnel identification samples, understanding which personnel have participated in abnormal behaviors and the specific characteristics and patterns of these abnormal behaviors, it provides an important reference basis for subsequent abnormal behavior detection and early warning.

[0054] Detailedly count the frequency of abnormal behaviors of the first person, which includes recording the number of times, frequency of abnormal behaviors of the first person, and the distribution of these behaviors in different time periods, etc. Through the analysis of these data, understand the abnormal behavior patterns of the first person, such as whether there are specific time rules, whether they are related to certain specific events or situations, etc. This statistical analysis can provide an important basis for subsequent abnormal behavior detection, judge the possibility of the first person having abnormal behaviors again according to the frequency characteristics of the first abnormal behaviors, and accordingly adjust the degree of attention and warning level for him. At the same time, these characteristics also help the system discover potential abnormal behavior trends, take measures in advance for prevention and intervention, so as to improve the safety and stability of the campus.

[0055] The process of initializing the abnormal weight is to assign a value to the first person that can reflect his tendency of abnormal behaviors, and this value is the first abnormal weight. If the frequency of abnormal behaviors of the first person is high, then the corresponding first abnormal weight will increase accordingly, indicating that the possibility of this person having abnormal behaviors is greater and more attention needs to be given. On the contrary, if the frequency of abnormal behaviors of the first person is low, then his first abnormal weight will be relatively small. The first abnormal weight generated in this way will play an important role in subsequent abnormal probability calculation and warning generation. By comprehensively considering the first abnormal weights of each person and other relevant factors, more accurately judge whether there are abnormal situations on campus and issue warning information in a timely manner to ensure the safety and order of the campus.

[0056] When an object enters the first monitoring blind area, real-time monitoring and analysis are carried out. If the objects entering this blind area include individuals previously marked as the first person, the first abnormal weight corresponding to this first person will be added to the initialized abnormal weight set. By adding the first abnormal weight to the set, more comprehensively consider the historical abnormal behavior situations of the people entering the blind area. The first abnormal weight reflects the possibility or tendency degree of the first person having abnormal behaviors. Adding it to the set can more accurately evaluate the abnormal degree of the current situation in subsequent abnormal probability calculation. For example, if the first person has had abnormal behaviors many times in history and his first abnormal weight is high, then when he enters the first monitoring blind area, this situation will be paid more attention to because his presence may increase the risk of abnormal situations in the blind area. In this way, potential abnormal situations can be more effectively identified and warned, and the safety and management efficiency of the campus can be improved.

[0057] In a possible implementation manner, the first abnormal probability generation module 60 further includes:

[0058] If the object entering the first monitoring blind area does not appear in the historical campus abnormal behavior library, initialize the corresponding abnormal weight to zero.

[0059] Specifically, for the objects entering the first monitoring blind area, verification will be carried out to determine whether they have appeared in the historical campus abnormal behavior library. If an object has not appeared in this library, it means that there is no record of the object's previous abnormal behavior. In this case, in order to accurately evaluate the abnormal situation, the abnormal weights of the objects that have not appeared in the historical library are initialized. Since there is no historical data to support the possibility of the object having abnormal behavior, giving it a zero weight can avoid misjudgment of the abnormal probability. In this way, when calculating the first abnormal probability, it will not be disturbed by these objects without historical abnormal records, and can more accurately focus on those objects with potential abnormal risks, improving the accuracy and reliability of abnormal behavior detection.

[0060] In a possible implementation manner, the first abnormal probability generation module 60 further includes:

[0061] Obtain the non-zero weights in the initialized abnormal weight set; calculate the proportion coefficient of the non-zero weights; when the proportion coefficient is greater than the preset proportion coefficient threshold, superimpose the non-zero weights and then weight the initial abnormal probability, and when the proportion coefficient is less than or equal to the preset proportion coefficient threshold, weight the initial abnormal probability with the maximum weight value to generate the first abnormal probability.

[0062] Specifically, traverse the initialized abnormal weight set and filter out the elements with non-zero weight values. These non-zero weights represent that the corresponding monitored objects have abnormal behavior records in the historical campus abnormal behavior library. By obtaining these non-zero weights, the possibility of abnormal behavior in the current monitoring scenario can be more accurately evaluated, providing key information support for subsequent abnormal probability calculation and early warning decision-making.

[0063] Calculate the proportion coefficient of non-zero weights and compare it with the preset proportion coefficient threshold. If the proportion coefficient is greater than the preset proportion coefficient threshold, it indicates that the proportion of non-zero weights in the whole is relatively large. At this time, perform an overlay operation on these non-zero weights, and the result after overlay will be used to weight the initial anomaly probability. In this way, fully consider the influence of multiple objects with abnormal weights on the anomaly probability, so that the calculated first anomaly probability can more accurately reflect the actual situation. When the proportion coefficient is less than or equal to the preset proportion coefficient threshold, it means that the proportion of non-zero weights in the whole is relatively small. In this case, weight the initial anomaly probability with the maximum weight value. The maximum weight value is the largest weight value that appears in the initialized anomaly weight set. The purpose of doing this is to still be able to give a certain adjustment to the initial anomaly probability when the proportion of non-zero weights is small, to ensure the accuracy of the first anomaly probability. By selecting different weighting methods according to the size of the proportion coefficient, reasonably generate the first anomaly probability, thereby improving the accuracy and reliability of campus abnormal behavior detection.

[0064] It should be noted that the above-mentioned sequence of embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above description of specific embodiments of this specification has been made. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0065] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0066] This specification and the drawings are only exemplary descriptions of the present application and are considered to have covered any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. The campus abnormal behavior detection system based on artificial intelligence is characterized by: include: A first monitoring blind area determination module, the first monitoring blind area determination module is used to connect to the campus monitoring platform, determine a first monitoring blind area, and a first monitoring neighborhood connected to the first monitoring blind area; A first standard aggregation feature set establishment module, the first standard aggregation feature set establishment module is used to establish a first standard aggregation feature set of the first monitoring blind area, wherein the first standard aggregation feature set includes aggregation features of multiple time zones; A first monitoring video data acquisition module, the first monitoring video data acquisition module is used to acquire first monitoring video data monitored by the monitoring equipment in the first monitoring neighborhood; A first monitoring object timing generation module, which identifies and marks the persons who enter the first monitoring blind spot from the first monitoring neighborhood based on the first monitoring video data, and generates a first monitoring object timing; A first real-time aggregation feature generation module, which performs aggregation feature recognition based on the first monitored object time series to generate a first real-time aggregation feature; A first abnormal probability generation module, the first abnormal probability generation module is used to compare the first real-time aggregation feature with the first standard aggregation feature set, perform abnormal aggregation feature recognition, and generate a first abnormal probability; A first abnormal warning information generating module, the first abnormal warning information generating module is used to generate first abnormal warning information when the first abnormal probability is greater than a preset warning probability; The first abnormal probability generating module includes: Acquire a historical campus abnormal behavior library, wherein any abnormal behavior sample in the historical campus abnormal behavior library contains an abnormal person identification sample; Counting the first abnormal behavior frequency characteristics of any first person who has exhibited abnormal behavior based on the abnormal person identification sample; Initializing an abnormal weight for the first person based on the first abnormal behavior frequency feature configuration to generate a first abnormal weight; When the object entering the first monitoring blind spot includes the first person, the first abnormal weight is added to the initialized abnormal weight set.

2. The campus abnormal behavior detection system based on artificial intelligence as claimed in claim 1 is characterized in that: The first monitoring blind area determination module also includes: Taking the first monitoring blind area as the center, identifying the boundary monitoring device located in the first monitoring blind area; Acquiring monitoring parameters of the border monitoring device; A boundary monitoring range and a monitoring direction are determined based on the monitoring parameters to generate the first monitoring neighborhood.

3. The campus abnormal behavior detection system based on artificial intelligence as claimed in claim 1 is characterized in that: The first standard aggregation feature set establishment module also includes: Obtaining regular behavior samples of the first monitoring blind area in the multiple time zones; Extracting features based on the conventional behavior samples to generate clustering features of the multiple time zones, wherein the clustering features include a threshold value of the number of clustered personnel and a stable feature of a change rate of clustered personnel; The stable characteristic of the change rate of the number of people gathered represents the stability of the change rate of the number of people gathered.

4. The campus abnormal behavior detection system based on artificial intelligence as claimed in claim 1 is characterized in that: The first abnormal probability generating module also includes: Obtaining a first monitoring time zone of the first monitoring video data; matching a target standard aggregation feature in the first standard aggregation feature set according to the first monitoring time zone; The first real-time aggregation feature is compared with the matching target standard aggregation feature, an aggregation feature deviation is established, and the abnormal weight is fused in combination with the monitored object to generate the first abnormal probability.

5. The campus abnormal behavior detection system based on artificial intelligence as claimed in claim 4 is characterized in that: The first abnormal probability generating module also includes: Initializing abnormal weights for all objects in the first monitoring object time series that enter the first monitoring blind area to generate an initialized abnormal weight set; Identifying an initial abnormality probability based on the cluster feature deviation, wherein the initial abnormality probability is proportional to the magnitude of the cluster feature deviation; The initial abnormality probability is weightedly calculated using the initial abnormality weight set to generate the first abnormality probability.

6. The campus abnormal behavior detection system based on artificial intelligence as claimed in claim 1, characterized in that: If the object entering the first monitoring blind spot does not appear in the historical campus abnormal behavior library, the corresponding abnormal weight is initialized to zero.

7. The campus abnormal behavior detection system based on artificial intelligence as claimed in claim 6 is characterized in that: The first abnormal probability generating module also includes: Obtaining non-zero weights in the initialized abnormal weight set; A coefficient for the non-zero weight; When the proportion coefficient is greater than the preset proportion coefficient threshold, the initial abnormal probability is weighted after superimposing the non-zero weight; when the proportion coefficient is less than or equal to the preset proportion coefficient threshold, the initial abnormal probability is weighted with the maximum weight value to generate the first abnormal probability.

Images