Smart park safety management system

Abstract

The invention discloses a smart park safety management system, and relates to the technical field of smart parks, the system uses an environment sensor group and a positioning device to collect position, temperature, humidity, illumination intensity and other multi-dimensional data in real time through a behavior data collection and analysis module, and calculates the moving speed, path curvature and stay time; and constructing a behavior data set based on a geographic information system (GIS), and storing the behavior data set in a time sequence database. And the abnormal behavior detection and early warning module extracts a behavior data set, constructs a feature vector and calculates an abnormal behavior probability by using a machine learning algorithm. By setting an abnormal threshold value, the system can accurately detect behaviors such as border crossing, speed abnormity and path deviation, and triggers an alarm or prompt. According to the real-time analysis based on the behavior data set, the probability of false alarm and missing alarm in a traditional system is remarkably reduced, potential safety risks can be rapidly recognized, and protective measures can be taken in time.

Description

Technical Field

[0001] The present invention relates to the technical field of smart parks, and in particular to a smart park security management system. Background Art

[0002] With the popularization of smart park construction, park security management has gradually become an important part of the smart city system. The smart park security management system belongs to the application field of intelligent security monitoring technology. Its core goal is to improve the security protection level in the park through modern information technology such as the Internet of Things, artificial intelligence and geographic information systems. In this specific field, the "real-time virtual protection wall for the entire park" technology has become an important innovative means because of its invisible, dynamic and intelligent characteristics. This technology realizes accurate and efficient security monitoring of people, vehicles and objects in the park through dynamic adjustment of virtual boundaries, providing an intelligent solution for smart park security management.

[0003] At present, the security management of traditional parks mainly relies on physical fences, manual patrols and static monitoring equipment, which have significant limitations in the modern complex park environment. On the one hand, physical fences lack flexibility and cannot adapt to the characteristics of frequent personnel flow and diversified regional functions in the park; on the other hand, manual patrols and monitoring methods rely on manpower, which can easily lead to safety hazards due to untimely response or insufficient coverage. In addition, existing security management methods usually lack intelligent analysis and prediction capabilities based on real-time data, and cannot efficiently detect and warn abnormal behaviors. This situation often causes the park to be unable to respond quickly when faced with emergencies, reducing the overall level of security protection. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a smart park security management system that solves the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: including a virtual protection wall definition module, a behavior data collection and analysis module, an abnormal behavior detection and early warning module, a dynamic adjustment module and a risk prediction module;

[0006] The virtual protection wall definition module performs digital modeling according to the regional characteristics within the park and defines the virtual boundary;

[0007] The behavior data collection and analysis module collects multi-dimensional data in each area in real time through environmental sensors and positioning devices, calculates and outputs the moving speed and residence time of the object based on the multi-dimensional data, integrates it with the remaining behavior data to obtain a behavior data group, and constructs a time series database to store the behavior data group;

[0008] The abnormal behavior detection and early warning module constructs a feature vector based on the behavior data group, and uses a machine learning algorithm to calculate and output the abnormal behavior probability based on the feature vector, and sets an abnormal threshold and the abnormal behavior probability to perform an abnormal evaluation, and generates an alarm trigger mechanism based on the abnormal evaluation result;

[0009] The dynamic adjustment module extracts the feature vector, combines the environmental change amount, analyzes the virtual boundary change amount, and dynamically adjusts the boundary of the virtual protection wall of each area in the park based on the virtual boundary change amount to obtain a new virtual boundary;

[0010] The risk prediction module calculates and outputs a predicted risk level using a prediction model based on a historical behavior data group and a real-time behavior data group, performs risk assessment based on a set risk threshold and the predicted risk level, and generates a response mechanism based on the risk assessment result.

[0011] Preferably, the virtual protection wall definition module includes a campus layout modeling unit and a virtual protection wall definition unit;

[0012] The park layout modeling unit converts the physical layout data of the park into a digital map by using CAD and a geographic information system (GIS), and marks the spatial coordinate information of the physical layout data on the digital map. The space in the park is divided into multiple areas according to different security requirements, and the security levels are divided. Positioning equipment is installed on objects entering the park to collect the coordinate information of the objects in real time, and the boundary coordinates and the coordinate positions of the objects are displayed in the geographic information system (GIS).

[0013] The objects include people and vehicles;

[0014] The virtual protective wall definition unit preliminarily defines the virtual protective wall based on the boundary coordinates of the object device and the coordinate position of the object in the divided area, and calculates and outputs the virtual boundary of the object device through a boundary algorithm to define the position and range of the virtual protective wall.

[0015] Preferably, the behavior data collection and analysis module includes a multidimensional data collection unit, a multidimensional data processing unit and a behavior data storage unit;

[0016] The multi-dimensional data acquisition unit collects environmental data in the park area in real time by installing an environmental sensor group in the park; at the same time, based on the positioning device installed on the object, the position coordinates and residence time of each object are collected in real time;

[0017] The environmental data include temperature, humidity and light intensity;

[0018] The environmental sensor group includes a temperature sensor, a humidity sensor and a photosensor;

[0019] The collected environmental data, location coordinates and residence time are summarized to generate multi-dimensional data, which is then sent to the Geographic Information System (GIS) via wireless WIFI.

[0020] Preferably, the multidimensional data processing unit receives the multidimensional data in real time in a geographic information system (GIS), and performs calculations on the multidimensional data to analyze the behavior data of objects in the area, wherein the behavior data includes movement speed and path curvature;

[0021] The moving speed is obtained by calculating the moving distance and time difference of the object in the area;

[0022] The path curvature is obtained by combining the turning angle between two consecutive path segments when the object moves by detecting the movement of the object through the Geographic Information System (GIS), and calculating the curvature in combination with the moving distance of the object;

[0023] Then, the behavior data is aggregated with the environmental data, the location coordinates and the residence time to generate a behavior data group, and the behavior data group is normalized;

[0024] The data storage unit is used to construct a time series database, and set a write port, a write port, a real-time data table and a historical data table. The behavior data group acquired in real time is marked with a timestamp and written into the real-time data table in the time series database through the write port. The original behavior data group in the real-time data table is automatically transferred to the historical data table and sorted and stored in the order of timestamps.

[0025] Preferably, the abnormal behavior detection and early warning module includes an abnormal behavior analysis unit and an evaluation and early warning unit;

[0026] The abnormal behavior analysis unit extracts the moving speed, path curvature and dwell time in the behavior data group of the real-time data table through the write port of the time series database, and composes a feature vector. It also weights all the features in the feature vector through a machine learning algorithm and combines it with the Sigmoid function to calculate the probability of abnormal behavior and classify and identify each behavior of the object in the park area.

[0027] Preferably, the evaluation and early warning unit sets an abnormal threshold based on the security level of the virtual boundary, and performs abnormal evaluation with the abnormal behavior probability, analyzes the probability of abnormal behavior of objects in the area, and triggers an alarm based on the abnormal evaluation result. The specific evaluation content is as follows;

[0028] When the probability of abnormal behavior is greater than the abnormal threshold, it means that there is abnormal behavior in the virtual boundary of the current area. At this time, an alarm is triggered and a prompt is given in the geographic information system GIS;

[0029] When the abnormal behavior probability is less than or equal to the abnormal threshold, it means that there is no abnormal behavior of the object within the virtual boundary of the current area, and no alarm is triggered at this time to continue monitoring.

[0030] Preferably, the dynamic adjustment module includes a boundary change analysis module and a dynamic adjustment unit;

[0031] The boundary change analysis module prompts the user that the current area object has abnormal behavior when abnormal behavior is found in the abnormal assessment, and at this time, a prompt for processing or not processing is generated in the geographic information system GIS pop-up window;

[0032] If you choose to handle this prompt, inspect the abnormal area of ​​the park object;

[0033] If you choose not to process, the environmental data, location coordinates and residence time in the behavior data group are extracted through the write port, and the change data of the objects in the area between the current moment and the previous moment are calculated respectively. The change data includes the position change, residence time change and environmental adjustment change. Based on the change data, weighted calculation is performed to obtain the virtual boundary change.

[0034] Preferably, the dynamic adjustment unit introduces a dynamic update mechanism based on the virtual boundary change amount, calculates and outputs a new virtual boundary based on the current virtual boundary and the virtual boundary change amount through a dynamic virtual protective wall adjustment formula, and dynamically adjusts the boundary of the virtual protective wall.

[0035] Preferably, the risk prediction module includes a security risk prediction unit and a risk assessment response unit;

[0036] The safety risk prediction unit analyzes risk data by extracting behavioral data groups from the real-time data table and historical data table of the time series database. The risk data includes speed change rate, position change, residence time change and environmental adjustment change. The risk change data is summarized and weighted to obtain the risk change. The comprehensive safety risk value is then calculated and output through the safety risk prediction formula to analyze the safety hazards in the current area.

[0037] Preferably, the risk assessment response unit manually sets the risk threshold according to the security level divided by the user in the area, performs risk assessment with the comprehensive security risk value, determines the security status of the current area, and generates a response mechanism based on the risk assessment result. The specific assessment content is as follows;

[0038] When the comprehensive security risk value is greater than the risk threshold, it indicates that the security status of the current area is abnormal. At this time, the current area is marked as an abnormal risk area and the response mechanism is triggered;

[0039] When the comprehensive security risk value is ≤ the risk threshold, it means that the current regional security status is normal. At this time, there is no need to generate a response mechanism and continuous monitoring is required;

[0040] When the response mechanism is triggered again, the monitoring frequency of the current area is automatically increased by 50%, the access control is locked, the alarm is activated, and the security personnel are notified to check the current area environment.

[0041] The present invention provides a smart park security management system, which has the following beneficial effects:

[0042] (1) The system uses the virtual protective wall definition module to digitally model the park and define virtual boundaries. It combines the dynamic adjustment module to dynamically adjust the virtual protective walls of each area of ​​the park and obtain new virtual boundaries. The system can dynamically adjust the protection range based on real-time data analysis of environmental changes, location changes, and residence time changes, effectively adapt to the complex behaviors of people and objects in the park, and improve the protection capabilities of key areas. Compared with traditional static physical boundaries, virtual protective walls have greater flexibility and dynamic adaptability, and can significantly reduce the security risks of the park caused by the failure of fixed boundaries.

[0043] (2) The system uses the behavior data collection and analysis module. The system uses environmental sensors and positioning devices to collect multi-dimensional data such as location, temperature, humidity, and light intensity in real time, calculates the moving speed, path curvature, and dwell time, and constructs a behavior data group based on the geographic information system GIS and stores it in the time series database. The abnormal behavior detection and early warning module extracts the behavior data group, constructs feature vectors, and uses machine learning algorithms to calculate the probability of abnormal behavior. By setting abnormal thresholds, the system can accurately detect behaviors such as crossing the boundary, abnormal speed, and path deviation, and trigger alarms or prompts. This real-time analysis based on the behavior data group significantly reduces the probability of false alarms and missed alarms in traditional systems, can quickly identify potential safety risks, take protective measures in a timely manner, and effectively reduce the occurrence of safety hazards in the park.

[0044] (3) The system integrates real-time data and historical data through the time series database through the risk prediction module, calculates the risk change, outputs the comprehensive risk value, and evaluates it with the risk threshold. When the comprehensive risk value exceeds the risk threshold, the system marks the area as an abnormal risk area and triggers a response mechanism, including increasing the monitoring frequency, locking the access control, activating the alarm, and notifying security personnel. This module can not only perceive the security status of the area in real time, but also identify potential risks in advance through the prediction model, provide security managers with intelligent decision-making basis, effectively avoid property losses and personal injuries caused by emergencies, and comprehensively improve the safety prevention and control level of the park. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1This is a schematic diagram of the process of a smart park security management system of the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] Example 1

[0048] See also Figure 1 , the present invention provides a smart park security management system. To achieve the above purpose, the present invention is implemented through the following technical solutions: including a virtual protection wall definition module, a behavior data collection and analysis module, an abnormal behavior detection and early warning module, a dynamic adjustment module and a risk prediction module;

[0049] The virtual protection wall definition module performs digital modeling based on the regional characteristics of the park and defines the virtual boundary B k ;

[0050] The behavior data collection and analysis module collects multi-dimensional data in each area in real time through environmental sensors and positioning devices, and calculates and outputs the moving speed V and residence time Ts of the object based on the multi-dimensional data, and then integrates it with the remaining behavior data to obtain the behavior data group, and builds a time series database to store the behavior data group;

[0051] The abnormal behavior detection and early warning module constructs a feature vector F based on the behavior data group, and uses a machine learning algorithm to calculate and output the abnormal behavior probability A(F) based on the feature vector F, and sets the abnormal threshold F1 and the abnormal behavior probability A(F) for abnormal evaluation, and generates an alarm trigger mechanism based on the abnormal evaluation results;

[0052] The dynamic adjustment module extracts the feature vector F and combines it with the environmental change △E to analyze the virtual boundary change △B k (t), and based on the virtual boundary change △B k (t) Dynamically adjust the boundaries of the virtual protective walls of each area in the park to obtain a new virtual boundary newB k (t);

[0053] The risk prediction module uses the prediction model to calculate and output the predicted risk level R based on the historical behavior data set and the real-time behavior data set. k (t), when setting the risk threshold F2 and the predicted risk level R k(t) Conduct risk assessment and generate response mechanisms based on risk assessment results.

[0054] In this embodiment, the system defines a flexible and adjustable virtual boundary Bk through the virtual protective wall definition module by combining digital modeling with the characteristics of the park area, so that the protection range of the park can be accurately delineated and dynamically adjusted according to actual needs. The behavior data collection and analysis module collects multi-dimensional data in the park in real time by installing environmental sensors and positioning equipment, including behavioral data such as location, speed, and residence time. After data integration and time series database storage, it provides reliable basic data support for subsequent abnormal behavior detection and analysis. The abnormal behavior detection and early warning module uses machine learning algorithms to construct a feature vector F based on multi-dimensional data, and calculates the probability of abnormal behavior A(F) to monitor various behaviors in the park in real time. If there is a behavior that exceeds the predetermined abnormal threshold F1, an alarm is triggered immediately to help managers detect potential security threats in a timely manner. The dynamic adjustment module calculates the virtual boundary change △B according to the behavioral data characteristics of the environmental change △E k (t), and dynamically adjust the boundaries of the virtual protective walls of each area accordingly, so as to achieve the ability to flexibly respond to different security threats. Finally, the risk prediction module calculates the security risk level R in the area by analyzing historical data and real-time data using the prediction model. k (t), and compare it with the set risk threshold F2, timely evaluate and generate a response mechanism. This series of modules cooperate with each other to form a comprehensive security protection system. After implementing this system, the park can monitor and warn abnormal behaviors in real time, realize dynamic protection, and automatically respond when security risks are high, greatly improving the accuracy, flexibility and efficiency of the park's security management, ensuring the safe and stable operation of the park.

[0055] Example 2

[0056] This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the virtual protection wall definition module includes a ,campus layout modeling unit and a virtual protection wall definition unit;

[0057] The park layout modeling unit converts the physical layout data of the park into a digital map by using CAD and geographic information system GIS, and marks the spatial coordinate information of the physical layout data on the digital map. The space in the park is divided into multiple areas according to different security requirements, and the security levels are divided. Positioning equipment is installed on objects entering the park to collect the coordinate information of the objects in real time, and the boundary coordinates (x, y) and the coordinate position P (x, y) of the object are displayed in the geographic information system GIS.

[0058] Objects include people and vehicles;

[0059] The virtual protective wall definition unit preliminarily defines the virtual protective wall based on the boundary coordinates (x, y) of the object device in the divided area and the coordinate position P (x, y) of the object, and calculates and outputs the virtual boundary B of the object device through the boundary algorithm. k , define the location and range of the virtual protection wall;

[0060] Virtual Boundary B k The output is calculated by the following boundary algorithm;

[0061]

[0062] In the formula, B k represents the virtual protective wall boundary of each region k. Each region has a corresponding virtual protective wall boundary, called virtual boundary B k , such as data centers and parking areas, the boundaries of the virtual protective wall are not static, but can be dynamically adjusted according to actual conditions. It is a set within a region, consisting of multiple coordinate points (x, y). All points (x, y) in a region belong to the boundary of this virtual protective wall. If the coordinate position P(x, y) of an object falls within the boundary, the object device is considered to be in region k. (x, y) represents a location coordinate point in the park. The boundaries of all regions can be defined by coordinates. P(x, y) represents the coordinate position of an object device in the park at a certain time point t.

[0063] In this embodiment, the park layout modeling unit first uses CAD and geographic information system GIS to convert the physical layout of the park into a digital map, and on this basis, marks the spatial coordinate information. According to the different security requirements of the park, the system further divides the park into multiple areas and classifies them according to their security levels. This digital layout not only provides intuitive map support for security management, but also ensures that the security requirements of each area are effectively met. At the same time, the virtual protective wall definition unit preliminarily defines the boundary Bk of the virtual protective wall in each area by analyzing the boundary coordinates (x, y) of each area and the object coordinate position P (x, y) collected in real time, and combines the boundary algorithm to accurately calculate the virtual boundary. These virtual protective walls are not static, but are adjusted in real time according to the actual situation and the dynamic changes of the park, so that the system can flexibly respond to different security threats. The virtual protective wall of each area is composed of multiple coordinate points (x, y). When the coordinate position P (x, y) of an object falls into the virtual protective wall of a certain area, the object is considered to be in the area, and then the corresponding security monitoring and protection measures are enabled. The implementation of this module significantly improves the flexibility and accuracy of park security management. Through digital and intelligent regional division, each area in the park can be personalized according to actual needs. The protection configuration avoids the limitations of traditional static walls. The dynamically adjusted virtual boundary enables the system to respond to the mobility changes of people and vehicles in the park in real time, ensuring that the safety of each area is always under control. In addition, the use of positioning equipment and GIS technology enhances the real-time monitoring capability of all objects in the park, providing a solid foundation for subsequent abnormal behavior warning and risk prediction.

[0064] Example 3

[0065] This embodiment is explained in Example 2. Please refer to Figure 1 ,Specifically: the behavior data collection and analysis module includes a multi-dimensional data collection unit, a multi-dimensional data processing unit and a behavior data storage unit;

[0066] The multi-dimensional data acquisition unit collects environmental data in the park area in real time by installing an environmental sensor group in the park; at the same time, based on the positioning device installed on the object, the position coordinates P (x, y) and the residence time Ts of each object are collected in real time;

[0067] Environmental data include temperature T, humidity H and light intensity G;

[0068] The environmental sensor group includes a temperature sensor, a humidity sensor, and a light sensor;

[0069] The collected environmental data, location coordinates P (x, y) and residence time Ts are summarized to generate multi-dimensional data, which are sent to the geographic information system GIS via wireless WIFI.

[0070] The multi-dimensional data processing unit receives multi-dimensional data in real time in the geographic information system GIS, and performs calculation and processing on the multi-dimensional data to analyze the behavior data of the objects in the area, wherein the behavior data includes the moving speed V and the path curvature K;

[0071] The moving speed V is calculated by calculating the moving distance d of the object in the area and the time difference T t，t+n The specific algorithm formula is: Among them, T t，t+n represents the time difference from the tth moment to the t+nth moment, x1 represents the horizontal coordinate of the first position of the object, x2 represents the horizontal coordinate of the second position of the object, y1 represents the vertical coordinate of the first position of the object, and y2 represents the vertical coordinate of the second position of the object;

[0072] The path curvature K is detected by the Geographic Information System (GIS) when the object moves, and the turning angle O between two consecutive path segments is L，L+1 , and then calculate and obtain it in combination with the moving distance d of the object; the specific algorithm formula is: Among them, L，L+1 Indicates the turning angle between path L and the next path L+1;

[0073] Then, based on the behavior data and environmental data, the location coordinates P (x, y) and the residence time Ts, a behavior data group is generated, and then the behavior data group is normalized;

[0074] The data storage unit is used to build a time series database, and set a write port, a write port, a real-time data table and a historical data table. The behavior data group acquired in real time is marked with a timestamp and written into the real-time data table in the time series database through the write port. The original behavior data group in the real-time data table is automatically transferred to the historical data table and stored in sorted order according to the timestamp order.

[0075] In this embodiment, the system collects environmental data in real time by installing an environmental sensor group in the park through a multidimensional data acquisition unit, and obtains the precise position coordinates P (x, y) and residence time Ts of each object through the positioning device installed on the object. These data not only provide the spatial position of the object, but also reflect the real-time changes in the environment, forming a comprehensive data basis for the behavior of people and vehicles in the park. All collected environmental data, location coordinates and residence time information are summarized into multidimensional data and transmitted to the geographic information system GIS via wireless WiFi for further processing. Next, the multidimensional data processing unit receives and analyzes these data in real time in the geographic information system GIS to calculate the moving speed V and path curvature K of each object. Specifically, the moving speed is calculated by the displacement and time difference of the object in the area, and the path curvature K is obtained by detecting the angle change of the turn during the movement of the object. A complete behavior data group is formed, and the data group is normalized for subsequent analysis and modeling. Finally, the data storage unit constructs a time series database, and sets a real-time data table and a historical data table to ensure that all behavior data groups are accurately recorded and stored in order of timestamps. This module not only provides systematic storage and management of behavioral data, but also ensures timely updating of real-time data and efficient retrieval of historical data. The implementation of this module realizes all-round real-time monitoring of the behavior of objects in the park, and improves the intelligence and refinement of the park's security management. Through the multi-dimensional collection and analysis of environmental data, location data and behavioral data, the system can more accurately identify and predict potential risks and abnormal behaviors in the park. In addition, through the design of a time series database, the system can effectively trace historical data, conduct trend analysis and behavioral pattern recognition, thereby providing strong data support for subsequent anomaly detection, risk assessment and decision support.

[0076] Example 4

[0077] This embodiment is explained in Example 3, please refer to Figure 1 ,Specifically: the abnormal behavior detection and early warning module includes an abnormal ,behavior analysis unit and an evaluation and early warning unit;

[0078] The abnormal behavior analysis unit extracts the moving speed V, path curvature K and residence time Ts from the behavior data group of the real-time data table through the write port of the time series database, and forms a feature vector F. The expression of the feature vector F is F = [V, K, Ts]. The machine learning algorithm is used to weight all the features in the feature vector F and combine them with the Sigmoid function to calculate the abnormal behavior probability A(F), and classify and identify each behavior of the object in the park area.

[0079] The abnormal behavior probability A(F) is calculated and output by the following algorithm formula;

[0080]

[0081] In the formula, σ represents the Sigmoid function, which is used to map the weighted sum result to the range of [0,1], and β represents the probability. j represents the weight of the jth feature, F j Represents the jth feature in the feature vector, that is, F1=V, F2=K, F3=Ts.

[0082] The evaluation and early warning unit is based on the virtual boundary B k The abnormal threshold F1 is set based on the security level, and the abnormal behavior probability A(F) is used for abnormal evaluation. The probability of abnormal behavior of objects in the area is analyzed, and an alarm is triggered based on the abnormal evaluation results. The specific evaluation contents are as follows;

[0083] When the abnormal behavior probability A(F)>abnormal threshold F1, it means that the virtual boundary B of the current area k There are abnormal behaviors of objects in the area, such as objects moving too fast, objects moving on paths that deviate from the normal path, such as objects that have been away from the normal path for a long time, resulting in dense objects in the current area. At this time, an alarm is triggered and a prompt is given in the geographic information system GIS;

[0084] When the abnormal behavior probability A(F) ≤ abnormal threshold F1, it means that the virtual boundary B of the current area k If there is no abnormal behavior in the object, no alarm will be triggered and monitoring will continue.

[0085] In this embodiment, the system extracts the behavior data group in the real-time data table from the time series database through the abnormal behavior analysis unit, and forms a feature vector F = [V, K, Ts] with these data. Then, the machine learning algorithm is used to weight each feature in the feature vector, and the abnormal behavior probability A(F) is calculated in combination with the Sigmoid function to determine whether the behavior of the object in the park is abnormal. The Sigmoid function maps the weighted result to the range of [0,1] to obtain a probability value. The greater the probability of abnormal behavior, the more likely the behavior is abnormal. The evaluation and early warning unit sets the abnormal threshold F1 based on the security level of the virtual protective wall and compares it with the calculated abnormal behavior probability A(F). When the abnormal behavior probability A(F) is greater than the set abnormal threshold F1, the system considers that the behavior of the object is abnormal, such as moving too fast, deviating from the path, or staying too long in a place where it should not stay, triggering an alarm and displaying the alarm information in the GIS. Through the implementation of this module, abnormal behaviors in the park are efficiently monitored and identified, so that potential security threats can be responded to quickly. For example, if someone moves quickly or stays for a long time in a non-designated area, the system will immediately issue an alarm to remind the management personnel to take measures. This not only effectively improves the security protection capabilities of the park, but also reduces the workload of manual monitoring through automated abnormal behavior detection. The system's adaptive capabilities enable it to flexibly adjust thresholds based on the set security level and actual conditions, thereby ensuring the accuracy and timeliness of early warnings.

[0086] Example 5

[0087] This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the dynamic adjustment module includes a boundary change analysis ,module and a dynamic adjustment unit;

[0088] The boundary change analysis module prompts the user that the current area object has abnormal behavior when abnormal behavior is found in the abnormal assessment. At this time, the GIS pop-up window generates a prompt for processing or not processing;

[0089] If you choose to handle this prompt, inspect the abnormal area of ​​the park object;

[0090] If you choose not to process, then extract the environmental data, position coordinates P (x, y) and residence time Ts in the behavior data group through the write port, and calculate the change data of the object in the area between the current moment and the previous moment. The change data includes the position change △P, the residence time change △Ts and the environmental adjustment change △E. Then, based on the change data, weighted calculation is performed to obtain the virtual boundary change △B k (t);

[0091] The position change △P, the dwell time change △Ts and the environmental adjustment change △E are calculated by the following algorithm formula

[0092] △E=(T(t)-T(t-1))+(H(t)-H(t-1))+(G(t)-G(t-1))

[0093] Get: ; where T(t) represents

[0094] △Ts=Ts(t)-Ts(t-1)

[0095] The temperature at the current moment t, T(t-1) represents the temperature at the previous moment, H(t) represents the humidity at the current moment t, H(t-1) represents the humidity at the previous moment, G(t) represents the light intensity at the temperature at the current moment t, G(t-1) represents the light intensity at the previous moment, Ts(t) represents the residence time at the current moment t, Ts(t-1) represents the residence time at the previous moment, x t Indicates the horizontal axis coordinate of the object at the current time t, x t-1 Indicates the horizontal axis coordinate of the object at the previous moment, y t Represents the vertical axis coordinate of the object at the current time t, y t-1 Indicates the vertical coordinate of the object at the previous moment

[0096] Virtual boundary change △B k (t) Calculate the output using the following algorithm formula;

[0097] △B k (t)=a1·△P+a2·△Ts+a3·△E;

[0098] Wherein, a1, a2 and a3 represent the weight values ​​of the position change △P, the residence time change △Ts and the environmental adjustment change △E respectively, and their specific values ​​are set by the user.

[0099] The dynamic adjustment unit is based on the virtual boundary change △B k (t) Introduce a dynamic update mechanism based on the current virtual boundary B k and virtual boundary variation △B k (t), calculate and output the new virtual boundary newB through the dynamic virtual protection wall adjustment formula k (t), dynamically adjust the boundary of the virtual protection wall;

[0100] New virtual boundary newB k (t) adjusting the formula calculation output through a dynamic virtual protection wall;

[0101] newB k (t) = B k(t-1)+△B k (t).

[0102] In this embodiment, the system uses the boundary change analysis module to prompt the user that abnormal behavior may exist in the current area when abnormal behavior is detected, and provides the option of processing or not processing. If processing is selected, the system will automatically trigger the inspection process; if not processing is selected, the change data of the object in the area between the current moment and the previous moment is calculated by extracting the environmental data, position coordinates P (x, y) and residence time Ts in the behavior data group, including the position change △P, residence time change △Ts and environmental adjustment change △E. These change data will be used to obtain the virtual boundary change △B through weighted calculation k (t), and use it as the basis for dynamic adjustment. The change includes the weight of location, residence time and environmental factors. The virtual boundary change reflects the behavioral characteristics and environmental changes of objects in the park, so that the position of the virtual protective wall can be adjusted more accurately. In the dynamic adjustment unit, based on the virtual boundary change △B k (t), the system will calculate the new virtual boundary newB through the dynamic virtual protection wall adjustment formula k (t). This new virtual boundary will dynamically adjust the security boundary within the park, so that the protective wall can respond to changes in the environment and behavior within the park in real time, thereby responding to security threats more flexibly and intelligently. Through the implementation of this module, the park's virtual protective wall can automatically adjust according to the actual environment and behavior data, enhancing the park's dynamic security protection capabilities. For example, when it rains and the humidity in a certain area is too high, causing the people in the area to run faster and abnormal behavior occurs, the unit system will automatically adjust the virtual boundary of the area according to the change data to ensure that the range of the protective wall adapts to new security requirements in a timely manner. This not only improves the security of the park, but also enhances the adaptability and intelligence level of the system, reduces manual intervention, and improves the management efficiency and response speed of the park. Through this dynamic adjustment, the park can maintain a flexible protection strategy in a constantly changing environment, effectively reduce potential safety hazards, and ensure the safety of people and property in the park.

[0103] Example 6

[0104] This embodiment is explained in Example 5. Please refer to Figure 1 ,Specifically: the risk prediction module includes a security risk prediction unit and a risk ,assessment response unit;

[0105] The safety risk prediction unit extracts the behavior data groups in the real-time data table and historical data table of the time series database to analyze the risk data, which includes the speed change rate △V, the position change △P, the residence time change △Ts and the environmental adjustment change △E, and summarizes the risk change data by weighted calculation to obtain the risk change △R k (t), and then calculate and output the comprehensive safety risk value R through the safety risk prediction formula k (t), analyze the safety risks in the current area;

[0106] Comprehensive safety risk value R k (t) Calculate the output through the following security risk prediction formula;

[0107] R k (t) = R k (t-1)+△R k (t);

[0108] Among them, the risk change △R k (t) The specific algorithm formula is: △R k (t) = △B k (t)+a4·△V, a4 represents the weight value of the speed change rate △V, and its specific value is set by the user.

[0109] The risk assessment response unit manually sets the risk threshold F2 based on the security level divided by the user according to the area, and compares it with the comprehensive security risk value R k (t) Conduct risk assessment to determine the safety status of the current area and generate a response mechanism based on the risk assessment results. The specific assessment contents are as follows;

[0110] When the comprehensive safety risk value R k When (t)>risk threshold F2, it indicates that the security status of the current area is abnormal. At this time, the current area is marked as an abnormal risk area and the response mechanism is triggered;

[0111] When the comprehensive safety risk value R k When (t)≤risk threshold F2, it means that the current regional security status is normal. At this time, there is no need to generate a response mechanism and continuous monitoring is required;

[0112] When the response mechanism is triggered again, the monitoring frequency of the current area will be automatically increased by 50%, the access control will be locked, the alarm will be activated, and the security personnel will be notified to check the current area environment.

[0113] In this embodiment, the implementation method of this module first relies on the real-time data and historical data in the time series database to extract key behavior data groups, including speed change rate △V, position change △P, residence time change △Ts and environmental adjustment change △E. After these data are aggregated and weighted, the risk change △R is obtained. k (t), and calculate the comprehensive safety risk value R through the safety risk prediction formula k (t) is used to assess the safety hazards in the current area. During this process, users can manually set various risk factors to ensure that the risk prediction matches the actual safety requirements of the park. Based on the calculated comprehensive safety risk value R k (t), the system can judge the safety status of the area and make response decisions. The risk assessment response unit performs risk assessment based on the set risk threshold F2. k (t) exceeds the risk threshold F2, indicating that the security status of the current area is abnormal. The system will mark the area as an abnormal risk area and trigger the response mechanism. The response mechanism includes automatically increasing the monitoring frequency of the area by 50%, locking the access control, starting the alarm system and notifying the security personnel to conduct on-site inspections. When the security risk is within an acceptable range, the system will continue to perform routine monitoring tasks. By implementing the risk prediction module, the park can evaluate security risks in real time and take proactive protective measures. This module significantly improves the intelligence level of park security management. It can not only quickly identify and respond to potential risks, but also effectively reduce human omissions or delayed responses. The enhanced monitoring frequency and immediate alarm mechanism enable security personnel to quickly locate and solve problems, greatly reducing the probability of security incidents and improving the overall security and emergency response efficiency of the park. In addition, through data-driven risk prediction and assessment, the system achieves more accurate risk control and optimizes the allocation and management efficiency of security resources in the park.

[0114] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A smart park security management system, characterized by: It includes virtual protection wall definition module, behavior data collection and analysis module, abnormal behavior detection and early warning module, dynamic adjustment module and risk prediction module; The virtual protection wall definition module performs digital modeling according to the regional characteristics within the park and defines the virtual boundary; The behavior data collection and analysis module collects multi-dimensional data in each area in real time through environmental sensors and positioning devices, calculates and outputs the moving speed and residence time of the object based on the multi-dimensional data, integrates it with the remaining behavior data to obtain a behavior data group, and constructs a time series database to store the behavior data group; The abnormal behavior detection and early warning module constructs a feature vector based on the behavior data group, and uses a machine learning algorithm to calculate and output the abnormal behavior probability based on the feature vector, and sets an abnormal threshold and the abnormal behavior probability to perform an abnormal evaluation, and generates an alarm trigger mechanism based on the abnormal evaluation result; The dynamic adjustment module extracts the feature vector, combines the environmental change amount, analyzes the virtual boundary change amount, and dynamically adjusts the boundary of the virtual protection wall of each area in the park based on the virtual boundary change amount to obtain a new virtual boundary; The risk prediction module calculates and outputs a predicted risk level using a prediction model based on a historical behavior data group and a real-time behavior data group, performs risk assessment based on a set risk threshold and the predicted risk level, and generates a response mechanism based on the risk assessment result.

2. The smart park security management system according to claim 1, characterized in that: The virtual protection wall definition module includes a campus layout modeling unit and a virtual protection wall definition unit; The park layout modeling unit converts the physical layout data of the park into a digital map by using CAD and a geographic information system (GIS), and marks the spatial coordinate information of the physical layout data on the digital map. The space in the park is divided into multiple areas according to different security requirements, and the security levels are divided. Positioning equipment is installed on objects entering the park to collect the coordinate information of the objects in real time, and the boundary coordinates and the coordinate positions of the objects are displayed in the geographic information system (GIS). The objects include people and vehicles; The virtual protective wall definition unit preliminarily defines the virtual protective wall based on the boundary coordinates of the object device and the coordinate position of the object in the divided area, and calculates and outputs the virtual boundary of the object device through a boundary algorithm to define the position and range of the virtual protective wall.

3. The smart park security management system according to claim 1 is characterized by: The behavior data collection and analysis module includes a multidimensional data collection unit, a multidimensional data processing unit and a behavior data storage unit; The multi-dimensional data acquisition unit collects environmental data in the park area in real time by installing an environmental sensor group in the park; at the same time, based on the positioning device installed on the object, the position coordinates and residence time of each object are collected in real time; The environmental data include temperature, humidity and light intensity; The environmental sensor group includes a temperature sensor, a humidity sensor and a photosensor; The collected environmental data, location coordinates and residence time are summarized to generate multi-dimensional data, which is then sent to the Geographic Information System (GIS) via wireless WIFI.

4. The smart park security management system according to claim 3 is characterized by: The multi-dimensional data processing unit receives the multi-dimensional data in real time in a geographic information system GIS, and performs calculations on the multi-dimensional data to analyze the behavior data of objects in the area, wherein the behavior data includes movement speed and path curvature; The moving speed is obtained by calculating the moving distance and time difference of the object in the area; The path curvature is obtained by combining the turning angle between two consecutive path segments when the object moves by detecting the movement of the object through the Geographic Information System (GIS), and calculating the curvature in combination with the moving distance of the object; Then, the behavior data is aggregated with the environmental data, the location coordinates and the residence time to generate a behavior data group, and the behavior data group is normalized; The data storage unit is used to construct a time series database, and set a write port, a write port, a real-time data table and a historical data table. The behavior data group acquired in real time is marked with a timestamp and written into the real-time data table in the time series database through the write port. The original behavior data group in the real-time data table is automatically transferred to the historical data table and sorted and stored in the order of timestamps.

5. The smart park security management system according to claim 4 is characterized in that: The abnormal behavior detection and early warning module includes an abnormal behavior analysis unit and an evaluation and early warning unit; The abnormal behavior analysis unit extracts the moving speed, path curvature and dwell time in the behavior data group of the real-time data table through the write port of the time series database, and composes a feature vector. It also weights all the features in the feature vector F through a machine learning algorithm and combines it with the Sigmoid function to calculate the probability of abnormal behavior and classify and identify each behavior of the object in the park area.

6. The smart park security management system according to claim 5, characterized in that: The evaluation and early warning unit sets an abnormal threshold based on the security level of the virtual boundary, performs abnormal evaluation based on the abnormal behavior probability, analyzes the probability of abnormal behavior of objects in the area, and triggers an alarm based on the abnormal evaluation result. The specific evaluation content is as follows; When the probability of abnormal behavior is greater than the abnormal threshold, it means that there is abnormal behavior in the virtual boundary of the current area. At this time, an alarm is triggered and a prompt is given in the geographic information system GIS; When the abnormal behavior probability is less than or equal to the abnormal threshold, it means that there is no abnormal behavior of the object within the virtual boundary of the current area, and no alarm is triggered at this time to continue monitoring.

7. The smart park security management system according to claim 4 is characterized by: The dynamic adjustment module includes a boundary change analysis module and a dynamic adjustment unit; The boundary change analysis module prompts the user that the current area object has abnormal behavior when abnormal behavior is found in the abnormal assessment, and at this time, a prompt for processing or not processing is generated in the geographic information system GIS pop-up window; If you choose to handle this prompt, inspect the abnormal area of ​​the park object; If you choose not to process, the environmental data, location coordinates and residence time in the behavior data group are extracted through the write port, and the change data of the objects in the area between the current moment and the previous moment are calculated respectively. The change data includes the position change, residence time change and environmental adjustment change. Based on the change data, weighted calculation is performed to obtain the virtual boundary change.

8. The smart park security management system according to claim 7, characterized in that: The dynamic adjustment unit introduces a dynamic update mechanism based on the virtual boundary variation, calculates and outputs a new virtual boundary based on the current virtual boundary and the virtual boundary variation through a dynamic virtual protection wall adjustment formula, and dynamically adjusts the boundary of the virtual protection wall.

9. The smart park security management system according to claim 8, characterized in that: The risk prediction module includes a security risk prediction unit and a risk assessment response unit; The safety risk prediction unit analyzes risk data by extracting behavioral data groups from the real-time data table and historical data table of the time series database. The risk data includes speed change rate, position change, residence time change and environmental adjustment change. The risk change data is summarized and weighted to obtain the risk change. The comprehensive safety risk value is then calculated and output through the safety risk prediction formula to analyze the safety hazards in the current area.

10. The smart park security management system according to claim 9, characterized in that: The risk assessment response unit manually sets the risk threshold according to the security level divided by the user according to the area, performs risk assessment with the comprehensive security risk value, determines the security status of the current area, and generates a response mechanism based on the risk assessment result. The specific assessment content is as follows; When the comprehensive security risk value is greater than the risk threshold, it indicates that the security status of the current area is abnormal. At this time, the current area is marked as an abnormal risk area and the response mechanism is triggered; When the comprehensive security risk value is ≤ the risk threshold, it means that the current regional security status is normal. At this time, there is no need to generate a response mechanism and continuous monitoring is required; When the response mechanism is triggered again, the monitoring frequency of the current area is automatically increased by 50%, the access control is locked, the alarm is activated, and the security personnel are notified to check the current area environment.

Images