Campus safety monitoring method and system for smart campus based on Internet of Things

Abstract

The invention relates to the technical field of intelligent monitoring, in particular to a campus safety monitoring method and system for an intelligent campus based on the Internet of Things, and the method comprises the steps: collecting video monitoring data of each region of a school, calculating a people flow dynamic change entropy value of the data of each video monitoring region in a collection process, and carrying out the calculation of the entropy value; and dividing a key people flow area and a non-key people flow area in the school according to the entropy value, setting an acquisition frequency for the key people flow area, and acquiring the image frames when people movement is detected in the non-key people flow area. According to the invention, the system can intelligently distinguish key people flow areas and non-key people flow areas in the campus by dynamically collecting video monitoring data and calculating the people flow dynamic change entropy. For these areas, the system can dynamically adjust the video monitoring acquisition frequency, performs higher-frequency acquisition in a key area, and only acquires data when people move in a non-key area, thereby optimizing the use of monitoring resources and data storage.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent monitoring, and particularly to a campus security monitoring method and system for smart campuses based on the Internet of Things. Background Art

[0002] The technical field of intelligent monitoring includes a technical system that uses sensors, cameras, and data communication technologies to monitor the environment in real time, collect data, and perform analysis, thereby realizing functions such as security management, behavior analysis, and data recording. The core content of this field involves interconnecting multiple monitoring devices and sensors through Internet of Things technology to achieve real-time transmission and processing of information. Intelligent monitoring is not only applied to home and public security, but also widely used in environments such as schools, hospitals, and enterprises, aiming to improve management efficiency and security. The core technologies include image recognition, sensor data processing, data fusion, cloud computing, and edge computing, ensuring that the system can efficiently and accurately capture and analyze monitoring data and respond in a timely manner.

[0003] Among them, campus security monitoring for smart campuses based on the Internet of Things refers to a security monitoring system specifically designed for the campus environment, which uses Internet of Things technology for device interconnection and real-time collection of various data including video monitoring, environmental sensor data, access control systems, etc. This patent theme addresses multiple technical matters in campus security monitoring, including how to connect devices such as cameras and sensors to the central system through the network, how to collect and transmit this data, and how to monitor and process various types of information collected by sensors and cameras in real time. Specifically, through the deployment and information interaction of Internet of Things devices, it realizes the monitoring and early warning of various security incidents on campus, and transmits the information to the management end through a data transmission channel for analysis and processing.

[0004] Existing monitoring systems usually rely on video data collection at a fixed frequency and fail to dynamically adjust the monitoring frequency and area according to the actual situation. Such systems handle the monitored area rather roughly and cannot intelligently distinguish key areas from non-key areas, resulting in a large amount of unnecessary data collection and transmission, which not only wastes storage space but also increases the pressure of data processing. Moreover, existing technologies usually lack detailed processing of real-time behavior analysis. Although some abnormal situations can be detected, it is often difficult to predict and early warn the occurrence of abnormal behaviors in a timely manner. Existing technologies also have insufficient prediction of the movement paths of personnel, resulting in a lack of sufficient foresight and countermeasures for possible security hazards. For example, when there is abnormal gathering or running of personnel, the system may not be able to promptly identify this behavior pattern and give accurate early warnings, possibly missing the opportunity to prevent at critical moments. These deficiencies make it difficult for existing technologies to meet the requirements of high efficiency and intelligence in the face of complex and changeable campus security management. Summary of the Invention

[0005] To solve the technical problems existing in the prior art, an embodiment of the present invention provides a campus security monitoring method and system for a smart campus based on the Internet of Things. The technical solutions are as follows:

[0006] A campus security monitoring method for a smart campus based on the Internet of Things includes the following steps:

[0007] S1: Collect video monitoring data of each area of the school, calculate the entropy value of the dynamic change of the number of people flowing in the data of each video monitoring area during the collection process, divide the key people flow areas and non-key people flow areas in the school according to the entropy value, set the collection frequency for the key people flow areas, and collect image frames when people movement is detected in the non-key people flow areas. Allocate weights according to the size of the entropy value and dynamically adjust the collection frequency of the monitoring screen to generate a dynamic monitoring data collection result;

[0008] S2: Based on the dynamic monitoring data collection result, perform clustering processing on the real-time obtained video monitoring images, and divide the categories of each behavior feature of the people in the school during the clustering process, including walking, gathering, and running, to generate a behavior feature classification result;

[0009] S3: Based on the behavior feature classification result, extract local feature points from the image frames corresponding to each type of behavior feature in the video monitoring, including the starting point, turning point, and ending point of the movement of the target people in the school, and refer to the local features to mark the people in the abnormal activity areas in the school to generate an abnormal person recognition result;

[0010] S4: Refer to the abnormal person recognition result, extract all the movement paths of the abnormal person in the school, predict the future movement path of the abnormal person, calculate the trajectory residual between the current movement path and the future movement path, identify the abnormal paths exceeding the preset residual threshold, and generate an abnormal person path detection result;

[0011] S5: According to the abnormal person path detection result, judge whether the abnormal person will continue to move towards the abnormal activity area in the future time period, and perform campus security early warning by means of alarms or real-time notifications to generate a campus security monitoring result.

[0012] The improvement of the present invention is that the dynamic monitoring data collection result includes regional entropy value, division of key and non-key people flow areas, setting of collection frequency, and adjustment of image frame collection; the behavior feature classification result includes behavior categories of walking, gathering, and running; the abnormal person recognition result includes local feature points, marking of abnormal activity areas, and abnormal behaviors of people; the abnormal person path detection result includes the current path, predicted path, and trajectory residual; the campus security monitoring result includes abnormal path warnings, personnel path abnormality reports, and real-time early warning notifications.

[0013] The improvement of the present invention is as follows: collecting video surveillance data of each area in the school, calculating the dynamic change entropy value of the data in each video surveillance area during the collection process, dividing the key pedestrian flow areas and non-key pedestrian flow areas in the school according to the entropy value, setting the collection frequency for the key pedestrian flow areas, collecting image frames when personnel movement is detected in the non-key pedestrian flow areas, allocating weights according to the size of the entropy value and dynamically adjusting the collection frequency of the surveillance images, and the specific steps for generating the dynamic surveillance data collection result are as follows:

[0014] S101: Collect video surveillance data of each area in the school, quantitatively analyze the personnel activity changes in the area by calculating the dynamic change entropy value of pedestrian flow, and generate the personnel activity change analysis result;

[0015] S102: Divide each area according to the personnel activity change analysis result, compare the size of the dynamic change entropy value of pedestrian flow with a preset entropy value threshold, judge the key pedestrian flow areas and non-key pedestrian flow areas, and generate the area division result;

[0016] S103: According to the area division result, set the collection frequency for the key pedestrian flow areas, conduct personnel movement monitoring for the non-key pedestrian flow areas, collect the image frames of the moving areas in real time, and dynamically adjust the collection frequency according to data fluctuations to generate the dynamic surveillance data collection result.

[0017] The improvement of the present invention is that for the dynamic change entropy value of each area, according to the formula:

[0018]

[0019] Calculate the dynamic change entropy value H(t);

[0020] where H(t) represents the dynamic entropy value at time t, p i (t) is the occurrence probability of the i-th behavior in the area at time t, and n represents the total number of all possible behaviors in the area;

[0021] Allocate weights according to the entropy value and adjust the data collection frequency, and the formula is as follows:

[0022]

[0023] where r(t) is the collection frequency at time t, w(t) is the weight calculated according to the dynamic change entropy value, and f(t) is an additionally introduced time-sensitive parameter.

[0024] The improvement of the present invention is that based on the dynamic surveillance data collection result, perform clustering processing on the real-time obtained video surveillance images, divide the categories of each behavior feature of the personnel in the school during the clustering process, including walking, gathering, running, and the specific steps for generating the behavior feature classification result are as follows:

[0025] S201: Extract the dynamic targets in the real-time video surveillance images of the dynamic monitoring data acquisition results, perform background segmentation, screen and mark the areas associated with personnel, and generate personnel area images;

[0026] S202: Analyze the motion features in the image according to the personnel area image, identify each behavior pattern, extract and calibrate the motion trajectories of each pattern, judge the behavior types of walking, gathering, and running, and generate behavior feature classifications;

[0027] S203: Classify each type of behavior according to the behavior feature classifications, obtain the occurrence frequency and duration of each type of behavior, and perform classification and calibration to generate behavior feature classification results.

[0028] The improvement of the present invention is that, based on the behavior feature classification results, local feature points are extracted from the image frames of the video surveillance corresponding to each type of behavior feature, including the moving start point, turning point, and moving end point of the target personnel on campus. Referring to the local features, the personnel located in the abnormal activity areas on campus are marked to generate the specific steps of the abnormal personnel recognition result as follows:

[0029] S301: Based on the behavior feature classification results, extract the motion trajectories of personnel from the video surveillance image frames corresponding to each type of behavior feature, identify the start point, turning point, and end point of each trajectory, and generate motion feature points;

[0030] S302: Based on the motion feature points, track the moving paths of personnel, analyze the start and end positions of the paths, and calibrate the positions of the abnormal activity areas. The abnormal activity areas include the wall area, warehouse area, closed area, and the area where power equipment is located, and generate abnormal activity area marks;

[0031] S303: According to the abnormal activity area marks, referring to the motion feature points, identify and mark the personnel located in the abnormal areas, complete the calibration and recording of the abnormal personnel, and generate the abnormal personnel recognition result.

[0032] The improvement of the present invention is that, referring to the abnormal personnel recognition result, extract all the moving paths of the abnormal personnel on campus, predict the future moving paths of the abnormal personnel, calculate the trajectory residuals between the current moving path and the future moving path, and identify the abnormal paths exceeding the preset residual threshold to generate the specific steps of the abnormal personnel path detection result as follows:

[0033] S401: Based on the abnormal personnel recognition result, extract the moving paths of the abnormal personnel on campus, record the node positions of each path segment, obtain the time interval and spatial change of each path node, and generate the abnormal personnel moving path;

[0034] S402: Predict future trajectory changes based on the current abnormal personnel movement path, with reference to the current movement mode and video surveillance image environment, to generate a prediction result of the future movement path;

[0035] S403: Calculate the trajectory residuals between the current path and the predicted path based on the prediction result of the future movement path, analyze the path differences, identify abnormal paths exceeding the preset threshold, and generate a detection result of the abnormal personnel path.

[0036] The improvement of the present invention is that for the trajectory residuals between the predicted trajectory and the actual trajectory, the formula is adopted:

[0037]

[0038] Calculate the trajectory residual Δx(t) between the predicted trajectory and the actual trajectory;

[0039] where Δx(t) represents the trajectory residual at time t, x p (t) and y p (t) are the horizontal and vertical coordinates of the predicted trajectory at time t, x a (t) and y a (t) are the horizontal and vertical coordinates of the actual trajectory at time t, λ(t) is an adjustment coefficient related to time, and σ is the dynamic uncertainty parameter of the trajectory point.

[0040] The improvement of the present invention is that according to the detection result of the abnormal personnel path, it is judged whether the abnormal personnel will continue to move towards the abnormal activity area in the future time period, and campus security early warning is carried out by means of alarm or real-time notification, and the specific steps for generating the campus security monitoring result are as follows:

[0041] S501: Based on the detection result of the abnormal personnel path, analyze the current path of the abnormal personnel and the abnormal activity area, and judge whether to continue approaching the target area according to the distance and moving direction between the path and the target area, and generate a path movement judgment result;

[0042] S502: Based on the path movement judgment result, evaluate the activities of the abnormal personnel in the future time period, calculate the time when the abnormal personnel arrives at the abnormal activity area and infer the change trend of the future path of the abnormal personnel, and generate a future movement evaluation result;

[0043] S503: According to the future movement evaluation result, trigger an alarm or real-time notification to send a warning message to the security management personnel, and generate a campus security monitoring result.

[0044] A campus security monitoring system for an intelligent campus based on the Internet of Things, the system includes:

[0045] The data acquisition module collects the video surveillance data of each area of the school, calculates the dynamic change entropy value of the data in each area, divides the areas within the school into key pedestrian flow areas and non-key pedestrian flow areas according to the entropy value, sets the acquisition frequencies of different areas, collects image frames through the movement of personnel in the non-key pedestrian flow areas, and allocates weights according to the dynamic change entropy value to dynamically adjust the acquisition frequency, generating the acquisition result of dynamic surveillance data;

[0046] Based on the acquisition result of the dynamic surveillance data, the behavior feature recognition module clusters the real-time video images, recognizes the behavior features of the personnel within the school, and divides them into behavior types of walking, gathering, and running. By analyzing the categories of each behavior feature, the behavior feature classification result is generated;

[0047] Based on the behavior feature classification result, the abnormal personnel recognition module extracts local feature points from the image frames of the video surveillance, recognizes the starting point, turning point, and ending point of the movement of the target personnel, calibrates the abnormal activity area, and conducts the recognition of abnormal personnel based on the calibration information, generating the recognition result of abnormal personnel;

[0048] Based on the recognition result of the abnormal personnel, the path detection module extracts the complete movement path of the abnormal personnel within the campus, uses the relevant data of the movement trajectory and behavior pattern for path prediction, calculates the trajectory residual between the current path and the predicted path, recognizes the abnormal path exceeding the preset residual threshold, and generates the path detection result of the abnormal personnel;

[0049] Based on the path detection result of the abnormal personnel, the security warning module determines whether the abnormal personnel continue to move towards the abnormal activity area, issues an alarm or notification through the real-time monitoring and alarm system, and notifies the campus security management personnel in real time, generating the campus security monitoring result.

[0050] The beneficial effects brought by the technical solution provided by the embodiment of the present invention at least include:

[0051] In the present invention, by dynamically collecting video surveillance data and calculating the entropy value of the dynamic change of the flow of people, the system can intelligently distinguish the key flow areas and non-key flow areas on campus. For these areas, the system can dynamically adjust the video surveillance collection frequency, collecting data at a higher frequency in key areas and only collecting data when people are moving in non-key areas, thereby optimizing the use of surveillance resources and data storage. This dynamic adjustment mechanism greatly improves the surveillance efficiency, reduces unnecessary redundant data storage and transmission, and at the same time ensures that key areas can be monitored in a timely and comprehensive manner. By performing cluster analysis on behavioral characteristics, the system can identify different types of activities, such as walking, gathering, running, etc., and accurately identify abnormal behaviors. By extracting key points such as the starting point of movement, turning points, and ending points, and combining with future path prediction, the system can identify potential abnormal behaviors or abnormal personnel paths in advance and issue early warnings in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0053] Figure 1 It is the flowchart of the method of the present invention;

[0054] Figure 2 It is the schematic diagram of the detailed process of step S1 of the present invention;

[0055] Figure 3 It is the schematic diagram of the detailed process of step S2 of the present invention;

[0056] Figure 4 It is the schematic diagram of the detailed process of step S3 of the present invention;

[0057] Figure 5 It is the schematic diagram of the detailed process of step S4 of the present invention;

[0058] Figure 6 It is the schematic diagram of the detailed process of step S5 of the present invention;

[0059] Figure 7 It is the system module diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] The following will describe the technical solutions in the present invention with reference to the drawings.

[0061] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as an "example" in the present invention should not be construed as being more preferred or more advantageous than other embodiments or design solutions. Rather, the use of the word "example" is intended to present concepts in a specific manner. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one of the two.

[0062] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the meaning they express is the same. "Of", "corresponding", and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the meaning they express is the same.

[0063] In the embodiments of the present invention, sometimes a subscript such as W 1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0064] To make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0065] Please refer to Figure 1 , the embodiments of the present invention provide a campus security monitoring method for a smart campus based on the Internet of Things, including the following steps:

[0066] S1: Collect video monitoring data of each area of the school, calculate the entropy value of the dynamic change of the flow of people in the data of each video monitoring area during the collection process, divide the key flow areas and non-key flow areas in the school according to the entropy value, set the collection frequency for the key flow areas, collect image frames when personnel movement is detected in the non-key flow areas, allocate weights according to the size of the entropy value, and dynamically adjust the collection frequency of the monitoring screen to generate a dynamic monitoring data collection result;

[0067] S2: Based on the dynamic monitoring data collection result, perform clustering processing on the video monitoring images obtained in real time, and divide the categories of each behavior feature of the personnel in the school during the clustering process, including walking, gathering, and running, to generate a behavior feature classification result;

[0068] S3: Based on the behavior feature classification result, extract local feature points from the image frames corresponding to each type of behavior feature in the video monitoring, including the starting point, turning point, and ending point of the movement of the target personnel in the school, and refer to the local features to mark the personnel in the abnormal activity areas in the school to generate an abnormal personnel recognition result;

[0069] S4: Refer to the identification results of abnormal personnel, extract all the movement paths of abnormal personnel within the school, predict the future movement paths of abnormal personnel, calculate the trajectory residuals between the current movement path and the future movement path, identify abnormal paths exceeding the preset residual threshold, and generate the detection results of abnormal personnel paths;

[0070] S5: According to the detection results of abnormal personnel paths, determine whether the abnormal personnel will continue to move towards the abnormal activity area in the future time period, and conduct campus security early warning through alarms or real-time notifications, and generate campus security monitoring results.

[0071] The dynamic monitoring data collection results include regional entropy values, division of key and non-key pedestrian flow areas, collection frequency setting, image frame collection adjustment. The behavior feature classification results include walking, gathering, and running behavior categories. The abnormal personnel identification results include local feature points, marked abnormal activity areas, and abnormal behaviors of personnel. The abnormal personnel path detection results include the current path, predicted path, and trajectory residuals. The campus security monitoring results include abnormal path warnings, personnel path abnormality reports, and real-time early warning notifications.

[0072] Please refer to Figure 2 , collect the video monitoring data of each area of the school, calculate the dynamic change entropy value of the data in each video monitoring area during the collection process, divide the key pedestrian flow areas and non-key pedestrian flow areas within the school according to the entropy value, set the collection frequency for the key pedestrian flow areas, collect image frames when personnel movement is detected in the non-key pedestrian flow areas, allocate weights according to the size of the entropy value, and dynamically adjust the collection frequency of the monitoring screen. The specific steps for generating the dynamic monitoring data collection results are as follows:

[0073] S101: Collect the video monitoring data of each area of the school, conduct quantitative analysis on the personnel activity changes within the area by calculating the dynamic change entropy value of the personnel flow, and generate the analysis results of personnel activity changes;

[0074] Refer to the dynamic change entropy value of each area, according to the formula:

[0075]

[0076] Calculate the dynamic change entropy value H(t);

[0077] Among them, H(t) represents the dynamic entropy value at time t, indicating the randomness of the behaviors occurring within the area at time t. It is calculated by analyzing the occurrence probabilities of various behaviors in the video, specifically by calculating the probability distribution of all behaviors in the video and then obtaining the entropy value. The higher the entropy value, the more complex or uncertain the behaviors in the area. p i(t) is the occurrence probability of the i-th behavior in the region at time t, representing the occurrence probability of the i-th behavior at time t. This value is obtained by analyzing the surveillance video and classifying behaviors. For example, assume that within a certain period of time, the surveillance video analyzes 3 behaviors: stationary, moving, and abnormal behavior. The surveillance system will count the frequency of each behavior and use it as the occurrence probability of that behavior. n represents the total number of all possible behaviors in the region. It is the number of behavior classifications, which is determined according to the number of behavior categories set in the surveillance video analysis. For example, if there are 3 behavior classifications (such as stationary, walking, and running) in a region, then n = 3. The weight is assigned according to the entropy value and the data collection frequency is adjusted. The formula is as follows:

[0078]

[0079] Among them, r(t) is the collection frequency at time t, and w(t) is the weight calculated according to the dynamically changing entropy value, the weight calculated according to the dynamically changing entropy value. This weight reflects the magnitude of the adjustment impact on the video data collection frequency at that moment. It is achieved through a weighting strategy, and the weight is usually set according to the entropy value H(t). The specific calculation method is to dynamically adjust the weight of this region by comparing the entropy value of historical data and the entropy value of real-time data. For example, when the entropy value of the region is large, indicating that the region changes greatly, the weight w(t) should be increased to increase the data collection frequency. f(t) is an additional time-sensitive parameter introduced to adjust the flexibility of the collection frequency according to different time periods, the time-sensitive parameter. It adjusts the flexibility of the collection frequency according to specific time periods (such as day and night, weekdays and weekends, etc.). This parameter can be calculated based on the relationship between historical behavior data and time periods. For example, the system can analyze from historical data that there is a high probability of abnormal behavior in certain regions at night. Therefore, in the night period, f(t) can be set to a larger value to increase the data collection frequency. Through this formula, the collection frequency of the surveillance region can be automatically adjusted according to the change of the region entropy value and the influence of time, so as to optimize the video data collection process and improve the system efficiency.

[0080] For example, at a certain moment, the occurrence probabilities of behaviors in the region are: p 1 = 0.4 (stationary) p 2 = 0.3 (moving) p 3 = 0.3 (abnormal behavior), then calculate the entropy value at this moment:

[0081] H(t) = -[0.4log(0.4) + 0.3log(0.3) + 0.3log(0.3)] ≈ 1.029;

[0082] If the weight w(t) = 2 and the time-sensitive parameter f(t) = 1.5, then the collection frequency is calculated as:

[0083]

[0084] The results show that based on the comprehensive calculation of entropy value, weight, and time-sensitive parameters, the system can adjust the acquisition frequency to make the video acquisition process more efficient, and flexibly adjust the data acquisition frequency according to different time periods, improving the accuracy and real-time performance of monitoring.

[0085] S102: Divide each area according to the analysis result of personnel activity changes, compare the magnitude of the entropy value of the dynamic change of people flow with a preset entropy threshold value, judge the key people flow area and non-key people flow area, and generate an area division result;

[0086] The key criterion for area division is the entropy value of the dynamic change of each area. The entropy value of each area is obtained through the analysis and calculation of video monitoring data, reflecting the complexity of the changes in personnel behavior in that area. According to the entropy value and the preset threshold range, the key nature of the area is judged. For example, if the entropy value of a certain area is 1.5, and this value may be greater than the threshold value of 1.0, then this area is a "key people flow area". According to the monitoring data, the personnel behavior dynamics in these areas are relatively complex and change violently, and high-frequency data acquisition is required to ensure real-time performance; while the areas with lower entropy values are determined as "non-key people flow areas", and the data acquisition frequency can be reduced. The setting of the threshold value is obtained through the analysis of historical data and real-time acquisition data to ensure that the key areas and non-key areas can be distinguished.

[0087] S103: According to the area division result, set the acquisition frequency of the key people flow area, and for the non-key people flow area, conduct personnel movement monitoring, collect the image frames of the moving area in real time, and dynamically adjust the acquisition frequency according to data fluctuations to generate a dynamic monitoring data acquisition result;

[0088] According to the area division result, for each key people flow area, the system will set a higher acquisition frequency to ensure that the details of personnel behavior in the area can be captured in real time. These areas include places where personnel activities are intensive, such as classroom entrances, dormitory corridors, and campus squares. For non-key areas, such as parking lots or empty corridors in teaching buildings, the system will rely on personnel movement monitoring to collect activity data in real time and adjust the acquisition frequency according to data fluctuations. If the personnel activities in a certain area suddenly increase, such as during holidays or specific events, the system will automatically increase the data acquisition frequency of that area to meet the monitoring requirements for emergencies. The detection of data fluctuations usually depends on the comparative analysis with historical data. If the current personnel activity frequency is higher than a certain benchmark value, the acquisition frequency will increase accordingly, thus realizing a dynamically optimized monitoring data acquisition process.

[0089] Please refer to Figure 3, based on the results of dynamic monitoring data collection, perform clustering processing on the real-time video surveillance images obtained, classify each behavior characteristic category of the on-campus personnel during the clustering process, including walking, gathering, and running. The specific steps for generating the behavior characteristic classification results are as follows:

[0090] S201: Extract the dynamic targets in the real-time video surveillance images from the results of dynamic monitoring data collection, perform background segmentation, screen and mark the areas associated with personnel, and generate a personnel area image;

[0091] Extracting the dynamic targets in the real-time video surveillance images and performing background segmentation is the first step. The background segmentation technology processes the dynamic changes in each frame of the image and uses algorithms such as the frame difference method or background modeling based on the Gaussian Mixture Model (GMM) to separate the static background from the dynamic targets. The system determines which areas have larger dynamic targets based on the changing areas in each frame of the image, and determines whether the area belongs to the personnel activity area by calculating the activity intensity of these areas (such as the change in gray value or the amplitude of pixel point change). During this process, through preset thresholds, parameters such as the area and gray change amplitude of the dynamic target area will be used to screen the areas related to personnel activities. For example, when the change amplitude exceeds the set threshold, the system considers this area as a dynamic target area and may be related to personnel activities, thus generating a personnel area image. The screening of dynamic targets is based on preset criteria to ensure that the system accurately screens out the areas related to personnel activities during campus security monitoring.

[0092] S202: Analyze the motion characteristics in the image based on the personnel area image, identify each behavior pattern, extract the motion trajectory of each pattern and perform calibration, and judge the behavior types of walking, gathering, and running to generate a behavior characteristic classification;

[0093] Based on the personnel area image, the analysis of motion characteristics uses the optical flow method or object detection technology based on deep learning to identify the motion patterns in the image. By comparing the motion trajectories of the targets in consecutive frames of the image, the system calculates the motion characteristics of each behavior pattern and classifies them based on these characteristics. Specifically, the motion characteristics of the walking behavior include smaller displacement and relatively stable speed, the gathering behavior is manifested as multiple target points moving and staying in a specific area, and the running behavior is manifested as larger displacement and rapid changes. The system will calibrate different motion characteristics with behavior types and judge the behavior categories through preset speed and displacement thresholds. For example, when the displacement of an object is greater than the set step length threshold, the system can identify it as a running behavior; while a smaller displacement and slower speed are judged as a walking behavior. Through such analysis, the system can extract each type of behavior based on the motion trajectory and generate the corresponding behavior characteristic classification.

[0094] S203: Classify each type of behavior according to the behavior feature classification, obtain the occurrence frequency and duration of each type of behavior, and conduct classification calibration to generate the behavior feature classification result;

[0095] Based on the behavior feature classification, the system further analyzes the occurrence frequency and duration of each type of behavior. The frequency of a behavior is calculated by counting the number of occurrences of a certain behavior type within a specific time period, and the duration is obtained by calculating the difference between the start time and end time of each behavior. For example, the frequency of the walking behavior can be obtained by counting the number of walking trajectories within 24 hours, and the duration of the walking behavior is calculated based on the start and end times of each walking trajectory. After behavior classification, the system will calibrate the frequency and duration of each behavior and conduct further analysis based on these data. In actual monitoring, the system determines which behaviors belong to high-frequency and long-duration behaviors and which belong to low-frequency and short-duration behaviors by setting reasonable thresholds. The goal of this process is to provide basic data for the subsequent optimization and behavior prediction of the system, ensuring that the campus security monitoring system can achieve real-time monitoring and alarm triggering of key areas without increasing redundant data collection.

[0096] Please refer to Figure 4 , based on the behavior feature classification result, extract local feature points from the image frames corresponding to the video monitoring of each type of behavior feature, including the starting point, turning point, and ending point of the movement of the target person on campus, and refer to the local feature markers to identify the personnel in the abnormal activity area on campus. The specific steps for generating the abnormal personnel identification result are as follows:

[0097] S301: Based on the behavior feature classification result, extract the movement trajectories of personnel from the video monitoring image frames corresponding to each type of behavior feature, identify the starting point, turning point, and ending point of each trajectory, and generate movement feature points;

[0098] The process of extracting the video monitoring image frames corresponding to each type of behavior feature and generating movement feature points depends on the comparison of consecutive frame images. Through image processing techniques such as optical flow method and background modeling method, the system can identify and extract dynamic targets in the image. The trajectory of each dynamic target is obtained by calculating the movement changes between image frames, and the starting point, turning point, and ending point of the target's movement in space are judged. For the identification of movement trajectory points, the system will determine when to recognize them as starting points, turning points, or ending points based on the set speed threshold and movement range. For example, a pixel distance threshold is set, and when the target displacement exceeds this threshold, it is regarded as a new feature point. During this process, the system will track and calibrate each target according to the set maximum displacement threshold range within the area to generate corresponding movement feature points, ensuring the accurate identification of each movement trajectory.

[0099] S302: Based on the motion feature points, track the movement path of the person, analyze the starting and ending positions of the path, calibrate the positions of the abnormal activity areas, where the abnormal activity areas include the perimeter wall area, the warehouse area, the enclosed area, and the area where electrical equipment is located, and generate abnormal activity area markers.

[0100] Based on the motion feature points, the tracking of the person's movement path determines the starting and ending positions of the path by analyzing the distance and time differences between the motion feature points. The system determines the validity of the path by setting the maximum acceptable displacement range and time delay threshold for each path. For example, if the displacement of consecutive feature points on the path exceeds a predefined pixel threshold of 30 or the time interval exceeds 5 seconds, the system will determine that the path is interrupted and mark it as an abnormal path. For sensitive areas on campus, such as the perimeter wall area, warehouses, enclosed areas, etc., the system will judge whether these areas are abnormally accessed based on the starting and ending points of the path. When the path of the person overlaps with the positions of these sensitive areas, the system can identify this area as an abnormal activity area and mark it. According to the preset criteria, the definition of the abnormal area is based on the persistence and frequency of the person's movement trajectory within these areas. For example, when the appearance frequency of people in the area exceeds a certain number of times, it is regarded as abnormal behavior.

[0101] S303: According to the abnormal activity area markers, with reference to the motion feature points, identify and mark the people located within the abnormal areas, complete the calibration and recording of the abnormal people, and generate the abnormal person identification results.

[0102] According to the markers of the abnormal activity areas, the system will compare the motion feature points with the area calibration results to further identify and mark the people located within the abnormal areas. The system matches the trajectories of people in the real-time monitoring video with the pre-calibrated abnormal activity areas to confirm whether the people have entered the calibrated high-risk areas. This process determines whether it belongs to abnormal activity by matching the time and space ranges of the people's activity trajectories. For example, when the trajectory of a person enters an enclosed area and stays for more than 30 seconds, the system will mark this person as an abnormal person and generate the abnormal person identification results. Through continuous tracking of the movement trajectories and comparison with the area markers, the system can ensure the timely and accurate identification of abnormal people, thereby further improving the safety protection ability of the campus.

[0103] Please refer to Figure 5 , with reference to the abnormal person identification results, extract all the movement paths of the abnormal people on campus, predict the future movement paths of the abnormal people, calculate the trajectory residuals between the current movement path and the future movement path, identify the abnormal paths that exceed the preset residual threshold, and the specific steps for generating the abnormal person path detection results are as follows:

[0104] S401: Based on the recognition result of the abnormal person, extract the movement path of the abnormal person on campus, record the node positions of each section of the path, obtain the time interval and spatial change of each path node, and generate the movement path of the abnormal person.

[0105] When extracting the movement path of the abnormal person based on the recognition result of the abnormal person, the node positions of each section of the path will be recorded first. The time interval and spatial change of each path node are obtained by the system according to the real-time monitoring data. For example, the system will record the path step by step according to the set threshold (such as one node per second or every 10 meters of movement). The time interval of each section of the path is calculated by the time difference between adjacent nodes, usually using the inter-frame difference method to ensure that the position of each node can be accurately captured. In some cases, if the time interval between path nodes is too short (such as less than 1 second), it may be necessary to consider that the movement speed between nodes is relatively fast, and then adjust the acquisition frequency to ensure the coherence and accuracy of data acquisition. Through these technical means, the system can generate the complete movement path of the abnormal person and ensure the timeliness and spatial accuracy of the path information.

[0106] S402: According to the current movement path of the abnormal person, refer to the current movement mode and the video surveillance image environment, predict the future trajectory change, and generate the prediction result of the future movement path.

[0107] Based on the movement path of the abnormal person, the system combines the current movement mode and the video surveillance image environment to predict the future trajectory change. By analyzing the movement speed and movement direction of each node on the current path, the system can predict the future activities of the person. For example, when a person moves at a relatively fast speed on a certain path on campus, the system will predict the possible future movement trajectory of the person based on historical data and the current movement trend. If there is an abnormal change in a certain path, such as a sudden deceleration or a change in direction, the system will predict the possible future pause or path change by comparing the historical trajectory with the real-time data. During this process, the system dynamically adjusts the prediction result of the future path by setting a time window and a spatial range (for example, the time window is 5 minutes, and if the path change exceeds a certain set threshold, it is determined as a trajectory change). This prediction helps the campus security system to respond to potential risks or abnormal behaviors in real time and improve the real-time performance and accuracy of monitoring.

[0108] S403: According to the prediction result of the future movement path, calculate the trajectory residual between the current path and the predicted path, analyze the path difference, identify the abnormal path exceeding the preset threshold, and generate the detection result of the abnormal person's path.

[0109] Compare the trajectory prediction data with the movement trajectory of the abnormal behavior person collected, according to the formula:

[0110]

[0111] Calculate the trajectory residual Δx(t) between the predicted trajectory and the actual trajectory;

[0112] where Δx(t) represents the trajectory residual at time t, and x p (t) and y p (t) are the horizontal and vertical coordinates of the predicted trajectory at time t, and x a (t) and y a (t) are the horizontal and vertical coordinates of the actual trajectory at time t. λ(t) is an adjustment coefficient related to time, indicating the flexibility adjustment of trajectory prediction in different time periods. As the time period changes, the accuracy of trajectory prediction may vary. The dynamic adjustment coefficient is used to dynamically adjust the calculation weight, which is determined by monitoring the behavior pattern fluctuations in a specific time period. During the experiment, by analyzing the behavior data in different time periods, the adjustment coefficient corresponding to each time period is obtained. For example, in a certain time period when the personnel movement frequency is high, the value of λ(t) may be set to be larger. σ is the dynamic uncertainty parameter of the trajectory point, reflecting the stability and reliability of the trajectory prediction. By introducing the dynamic adjustment coefficient λ(t) and the uncertainty parameter σ, the calculation of the trajectory residual can be further optimized, and the differences in prediction accuracy in different time periods can be considered. x p (t) and y p (t) respectively represent the horizontal and vertical coordinates of the trajectory points predicted based on the movement pattern of abnormal behavior personnel, and x a (t) and y a (t) represent the horizontal and vertical coordinates of the actual trajectory points obtained through real-time monitoring data. By calculating the Euclidean distance between these points, the trajectory residual at each moment can be obtained. The introduction of the dynamic adjustment coefficient λ(t) is to dynamically adjust the calculation of the residual according to the data changes in the time period. Especially in different time periods (such as peak hours and off-peak hours), the accuracy of trajectory prediction may be different, so an adjustment coefficient is needed to ensure more accurate comparison. The introduction of the dynamic uncertainty parameter σ of the trajectory point is to consider that the reliability of trajectory prediction is different under different environmental conditions (such as light changes, occlusion, etc.), so the uncertainty parameter is used to increase the flexibility of the calculation.

[0113] For example, at a certain moment, the coordinates of the predicted trajectory point are (x p (5), y p (5)) = (15, 10), while the coordinates of the actual trajectory point are (x a (5), y a (5)) = (12, 9), and at this moment, the dynamic adjustment coefficient λ(5) = 1.2, and the uncertainty of the trajectory point σ = 0.5. Then the trajectory residual is:

[0114]

[0115] The results show that by introducing the dynamic adjustment coefficient and the uncertainty of trajectory points, the trajectory residuals can be calculated more accurately. Especially when facing the influence of different time periods or environmental factors, the flexibility and accuracy of trajectory prediction are enhanced. When the difference between the predicted trajectory and the actual trajectory reaches a certain threshold (such as more than 5 meters), it will be marked as an abnormal trajectory for further safety monitoring and early warning.

[0116] Please refer to Figure 6 , according to the detection results of abnormal personnel paths, judge whether the abnormal personnel will continue to move towards the abnormal activity area in the future time period, and conduct campus safety early warning through alarms or real-time notifications. The specific steps for generating campus safety monitoring results are as follows:

[0117] S501: Based on the detection results of abnormal personnel paths, analyze the current path of the abnormal personnel and the abnormal activity area. According to the distance and moving direction between the path and the target area, judge whether to continue approaching the target area and generate a path movement judgment result;

[0118] Based on the detection results of abnormal personnel paths, the system judges whether to continue approaching the target area by calculating the distance between the current path of the abnormal personnel and the target area and its moving direction. The system first uses real-time monitoring data to calculate the spatial distance between the current position of the personnel and the target area through the time interval and displacement of each path node. For example, when the personnel is less than 5 meters away from the target area, the system will further analyze whether they are moving towards the target area. If the current movement direction of the personnel is consistent with the target area and the path distance is gradually shrinking, the system judges that they are approaching the target area and generates a path movement judgment result. This process is judged according to the set threshold. For example, if the set radius of the target area is 10 meters, the system will set this range as the approaching area. When the path distance is less than the threshold and the direction is the same, it is determined that the personnel may enter the target area.

[0119] S502: Based on the path movement judgment result, evaluate the activities of the abnormal personnel in the future time period, estimate the time when the abnormal personnel reaches the abnormal activity area and speculate on the change trend of the future path of the abnormal personnel, and generate a future movement evaluation result;

[0120] Based on the path movement judgment result, the system evaluates the activities of abnormal personnel in the future time period. Through the analysis of real-time data and historical movement trajectories, the system can estimate the time when the abnormal personnel will reach the target area, and speculate on the future path change trend based on data such as their speed and acceleration. For example, if a person's moving speed in the past few minutes is 2 meters per second and there is no obvious deceleration, the system will predict that they will reach the target area in the next 5 minutes. The system will further monitor the movement state of the abnormal personnel. If their speed significantly slows down or there is a sudden stop, the system will adjust the prediction time accordingly and evaluate based on the current path change trend. When the path undergoes abnormal changes (such as a change in direction or a slowdown in speed), the system will update the expected arrival time in a timely manner and readjust the predicted path.

[0121] S503: According to the future movement evaluation result, through triggering an alarm or real-time notification, send a warning message to the security management personnel and generate the campus security monitoring result;

[0122] According to the future movement evaluation result, the system sends the warning message to the security management personnel through triggering an alarm or real-time notification. When the system predicts that the abnormal personnel may approach the target area and meet the preset conditions, it will automatically trigger an alarm. For example, if the distance between the person's path and the target area is less than 5 meters and it is expected to reach the target area within 5 minutes, the system will send an alarm notification to the security management personnel, prompting them to take corresponding emergency measures. In addition, the system will also adjust the alarm level according to the actual situation. If the proximity of the person's path to the target area gradually increases, or the arrival time at the target area is advanced, the alarm will be given priority and sent to the campus security management personnel through real-time notification to ensure timely response and handling of potential security threats.

[0123] Please refer to Figure 7 , the campus security monitoring system for smart campuses based on the Internet of Things, the system includes:

[0124] The data acquisition module collects video monitoring data of each area of the school, calculates the dynamic change entropy value of the data in each area, divides the school areas into key pedestrian flow areas and non-key pedestrian flow areas according to the entropy value size, sets the acquisition frequencies of different areas, the non-key pedestrian flow areas collect image frames through personnel movement, and assigns weights according to the dynamic change entropy value, dynamically adjusts the acquisition frequency, and generates the dynamic monitoring data acquisition result;

[0125] The behavior feature recognition module clusters the real-time video images based on the dynamic monitoring data acquisition result, recognizes the behavior features of the personnel in the school, divides them into behavior types of walking, gathering, and running, and generates the behavior feature classification result by analyzing the categories of each behavior feature;

[0126] Based on the classification results of behavioral characteristics, the abnormal personnel recognition module extracts local feature points from the image frames of video surveillance, identifies the starting point, turning point, and ending point of the target personnel's movement, calibrates the abnormal activity area, and conducts abnormal personnel recognition based on the calibration information to generate the abnormal personnel recognition result;

[0127] Based on the abnormal personnel recognition result, the path detection module extracts the complete movement path of the abnormal personnel on campus, uses the relevant data of the movement trajectory and behavior pattern for path prediction, calculates the trajectory residual between the current path and the predicted path, identifies the abnormal path exceeding the preset residual threshold, and generates the abnormal personnel path detection result;

[0128] Based on the abnormal personnel path detection result, the security warning module determines whether the abnormal personnel continues to move towards the abnormal activity area, and issues an alarm or notification through the real-time monitoring and alarm system to notify the campus security management personnel in real time, generating the campus security monitoring result.

[0129] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A campus security monitoring method for a smart campus based on the Internet of Things, characterized in that: The following steps are involved: S1: Collect video surveillance data from each area of ​​the school, calculate the entropy value of the dynamic change of the flow of people in each video surveillance area during the collection process, divide the key flow areas and non-key flow areas in the school according to the entropy value, set the collection frequency for the key flow areas, collect image frames when the movement of people is detected in the non-key flow areas, assign weights according to the entropy value and dynamically adjust the collection frequency of the monitoring screen to generate dynamic monitoring data collection results; S2: Based on the dynamic monitoring data collection results, cluster the video surveillance images acquired in real time, classify each behavior feature of the school personnel in the clustering process into categories, including walking, gathering, and running, and generate behavior feature classification results; S3: Based on the behavioral feature classification results, extract local feature points from the image frames of the video surveillance corresponding to each type of behavioral feature, including the moving starting point, turning point, and moving end point of the target person in the school, mark the people in the abnormal activity area in the school with reference to the local features, and generate abnormal person identification results; S4: referring to the abnormal person identification result, extracting all movement paths of the abnormal person in the school, predicting the future movement path of the abnormal person, calculating the trajectory residual between the current movement path and the future movement path, identifying the abnormal path exceeding the preset residual threshold, and generating the abnormal person path detection result; S5: Based on the abnormal person path detection result, determine whether the abnormal person will continue to move to the abnormal activity area in the future time period, and issue a campus safety warning through an alarm or real-time notification to generate a campus safety monitoring result.

2. The campus security monitoring method for a smart campus based on the Internet of Things according to claim 1 is characterized in that: The dynamic monitoring data collection results include regional entropy values, key and non-key human flow area divisions, collection frequency settings, and image frame collection adjustments. The behavior feature classification results include walking, gathering, and running behavior categories. The abnormal personnel identification results include local feature points, abnormal activity area marks, and abnormal personnel behavior. The abnormal personnel path detection results include current path, predicted path, and trajectory residuals. The campus safety monitoring results include abnormal path warnings, personnel path abnormality reports, and real-time early warning notifications.

3. The campus security monitoring method for a smart campus based on the Internet of Things according to claim 1 is characterized in that: Collect video surveillance data from each area of ​​the school, calculate the dynamic change entropy value of each video surveillance area data during the collection process, divide the key flow areas and non-key flow areas in the school according to the entropy value, set the collection frequency for the key flow areas, and collect image frames when the movement of people is detected in the non-key flow areas. According to the size of the entropy value, weights are assigned and the collection frequency of the monitoring screen is dynamically adjusted. The specific steps to generate dynamic monitoring data collection results are as follows: S101: Collect video surveillance data from each area of ​​the school, perform quantitative analysis on changes in personnel activities in the area by calculating entropy values ​​of dynamic changes in human flow, and generate analysis results of changes in personnel activities; S102: Divide each area according to the analysis result of the change of personnel activities, compare the size of the entropy value of the dynamic change of the flow of people with the preset entropy value threshold, determine the key flow of people area and the non-key flow of people area, and generate the area division result; S103: According to the area division result, the acquisition frequency of the key human flow area is set, and the movement of people in the non-key human flow area is monitored, and the image frames of the moving area are acquired in real time. The acquisition frequency is dynamically adjusted according to data fluctuations to generate dynamic monitoring data acquisition results.

4. The campus security monitoring method for a smart campus based on the Internet of Things according to claim 1 is characterized in that: For the dynamic change entropy value of the flow of people in each area, according to the formula: Calculate the dynamically changing entropy value H(t); Among them, H(t) represents the dynamic entropy value at time t, p i (t) is the probability of occurrence of the ith behavior in the region at time t, and n represents the total number of all possible behaviors in the region; The weights are allocated according to the entropy value and the data collection frequency is adjusted. The formula is as follows: Among them, r(t) is the acquisition frequency at time t, w(t) is the weight calculated according to the dynamically changing entropy value, and f(t) is an additional time-sensitive parameter.

5. The campus security monitoring method for a smart campus based on the Internet of Things according to claim 1 is characterized in that: Based on the dynamic monitoring data collection results, clustering processing is performed on the real-time acquired video surveillance images, and each behavior feature of the school personnel is classified into categories during the clustering process, including walking, gathering, and running. The specific steps for generating the behavior feature classification results are as follows: S201: extracting dynamic targets from the real-time video surveillance image in the dynamic surveillance data collection result, performing background segmentation, screening and marking areas associated with personnel, and generating a personnel area image; S202: Analyze the motion features in the image according to the personnel area image, identify each behavior pattern, extract and calibrate the motion trajectory of each pattern, determine the walking, gathering, and running behavior types, and generate a behavior feature classification; S203: Classify each type of behavior according to the behavior characteristics, obtain the occurrence frequency and duration of each type of behavior, and perform classification and calibration to generate a behavior characteristic classification result.

6. The campus security monitoring method for a smart campus based on the Internet of Things according to claim 1 is characterized in that: Based on the behavioral feature classification results, local feature points are extracted from the image frames of the video surveillance corresponding to each type of behavioral feature, including the moving starting point, turning point, and moving end point of the target person in the school. The people in the abnormal activity area in the school are marked with reference to the local features. The specific steps for generating abnormal person identification results are as follows: S301: Based on the behavior feature classification results, extract the movement trajectory of the person from the video surveillance image frame corresponding to each type of behavior feature, identify the starting point, turning point, and end point of each trajectory, and generate movement feature points; S302: Based on the motion feature points, the moving path of the personnel is tracked, the starting and ending positions of the path are analyzed, and the position of the abnormal activity area is calibrated. The abnormal activity area includes the wall area, the warehouse area, the closed area, and the area where the power equipment is located, and the abnormal activity area mark is generated; S303: According to the abnormal activity area mark, referring to the motion feature points, identifying and marking the persons in the abnormal area, completing the calibration and recording of the abnormal persons, and generating the abnormal person identification result.

7. The campus security monitoring method for a smart campus based on the Internet of Things according to claim 1 is characterized in that: Referring to the abnormal personnel identification results, extracting all movement paths of abnormal personnel in the school, predicting the future movement paths of abnormal personnel, calculating the trajectory residuals between the current movement paths and the future movement paths, identifying abnormal paths that exceed the preset residual threshold, and generating the abnormal personnel path detection results are as follows: S401: Based on the abnormal person identification result, extract the movement path of the abnormal person in the school, record the node position of each path, obtain the time interval and spatial change of each path node, and generate the movement path of the abnormal person; S402: predicting future trajectory changes based on the current movement path of the abnormal person, referring to the current movement pattern and the video surveillance image environment, and generating a future movement path prediction result; S403: According to the future moving path prediction result, the trajectory residual between the current path and the predicted path is calculated and the path difference is analyzed, abnormal paths exceeding a preset threshold are identified, and abnormal personnel path detection results are generated.

8. The campus security monitoring method for a smart campus based on the Internet of Things according to claim 1 is characterized in that: For the trajectory residual between the predicted trajectory and the current trajectory, the formula is used: Calculate the trajectory residual Δx(t) at time t; Among them, x p (t) and y p (t) is the horizontal and vertical coordinates of the predicted trajectory at time t, x a (t) and y a (t) is the horizontal and vertical coordinates of the actual trajectory at time t, λ(t) is the time-related adjustment coefficient, σ 2 is the dynamic uncertainty parameter of the trajectory point.

9. The campus security monitoring method for a smart campus based on the Internet of Things according to claim 1 is characterized in that: According to the abnormal person path detection result, it is determined whether the abnormal person will continue to move to the abnormal activity area in the future time period, and a campus safety warning is issued by means of an alarm or real-time notification. The specific steps for generating the campus safety monitoring result are as follows: S501: Based on the abnormal person path detection result, analyze the current path of the abnormal person and the abnormal activity area, determine whether to continue approaching the target area according to the distance between the path and the target area and the moving direction, and generate a path movement judgment result; S502: Based on the path movement judgment result, the activities of the abnormal person in the future time period are evaluated, the time when the abnormal person arrives at the abnormal activity area is estimated, and the change trend of the abnormal person's future path is inferred to generate a future movement evaluation result; S503: Based on the future mobility assessment results, an early warning message is sent to the security management personnel by triggering an alarm or a real-time notification, thereby generating a campus security monitoring result.

10. A campus security monitoring system for smart campuses based on the Internet of Things, characterized in that: According to any one of claims 1 to 9, the smart campus based on the Internet of Things is implemented using a campus security monitoring method, and the system comprises: The data acquisition module collects video surveillance data from each area of ​​the school, calculates the dynamic change entropy value of each area data, divides the school area into key flow areas and non-key flow areas according to the entropy value, sets the acquisition frequency of different areas, and collects image frames through the movement of people in non-key flow areas, and assigns weights according to the dynamically changing entropy value, dynamically adjusts the acquisition frequency, and generates dynamic monitoring data acquisition results; The behavior feature recognition module clusters the real-time video images based on the dynamic monitoring data collection results, identifies the behavior features of the people on campus, and divides them into walking, gathering, and running behavior types. By analyzing the category of each behavior feature, a behavior feature classification result is generated; The abnormal person identification module extracts local feature points from the image frames of the video surveillance based on the behavioral feature classification results, identifies the target person's movement starting point, turning point, and ending point, and calibrates the abnormal activity area, and performs abnormal person identification based on the calibration information to generate abnormal person identification results; The path detection module extracts the complete movement path of the abnormal person on campus based on the abnormal person identification result, uses the relevant data of the movement trajectory and behavior pattern to predict the path, calculates the trajectory residual between the current path and the predicted path, identifies the abnormal path that exceeds the preset residual threshold, and generates the abnormal person path detection result; The safety warning module determines whether the abnormal person continues to move to the abnormal activity area based on the abnormal person path detection results, issues an alarm or notification through the real-time monitoring and alarm system, and informs the campus security management personnel in real time to generate campus safety monitoring results.

Images