A smart early warning method and system for campus safety risks

Abstract

This invention discloses an intelligent early warning method and system for campus security risks. The invention relates to the technical field of intelligent early warning methods. It determines corresponding job identifiers based on the job information of various campus personnel and their corresponding personnel information; and establishes a campus intelligent management database based on the security levels of each activity area, the security matching coefficients of multiple adjacent activity areas, and deep learning of past campus security incidents, ensuring the accuracy of the campus intelligent management database. Therefore, in the campus intelligent management database, activity areas in a warning state are determined based on the analysis of campus early warning signals; and collaborative teams are determined based on the early warning events of the activity areas in a warning state, the security levels of surrounding activity areas, and the early warning matching table, thus improving the accuracy of intelligent early warning of campus security risks.

Description

Technical Field

[0001] This invention relates to the technical field of intelligent early warning methods, and more particularly to an intelligent early warning method and system for campus safety risks. Background Technology

[0002] With the development of technology, campus safety risks have gradually received attention. There are various personnel on campus, including students, teachers, administrative staff, environmental protection personnel, etc. Each person has corresponding job information. In the current technology, the safety status of various activity areas on campus is determined by one-way reporting from teachers or students. It is not possible to accurately control the safety level of each activity area or the safety status of multiple adjacent activity areas, resulting in low accuracy of campus safety risk early warning. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides an intelligent early warning method and system for campus security risks.

[0004] This invention provides an intelligent early warning method for campus security risks, comprising:

[0005] The corresponding job identifier is determined based on the job information of each campus staff member and the corresponding personnel information.

[0006] The activity areas of multiple campus personnel are determined based on the identification of each position, the three-dimensional model of the campus, and the time.

[0007] The security level of each activity area is determined based on its location, the type of activities within each area, and the density of people on campus.

[0008] The campus intelligent management database is determined based on the security level of each activity area, the security matching coefficient of multiple adjacent activity areas, and deep learning of past campus security incidents.

[0009] The campus intelligent management database identifies activity areas under warning status based on the analysis of campus warning signals.

[0010] The collaborative team is determined based on the warning events in the activity area under warning status, the security level of the surrounding activity areas, and the warning matching table.

[0011] This invention provides an intelligent early warning system for campus security risks, which is applied to the aforementioned intelligent early warning method for campus security risks. The intelligent early warning system for campus security risks includes:

[0012] The job identifier module is used to determine the corresponding job identifier based on the job information of each campus staff member and the corresponding personnel information.

[0013] The activity area module is used to determine the activity areas of multiple campus personnel based on their job identifiers, a 3D model of the campus, and the time.

[0014] The security level module is used to determine the security level of each activity area based on its location, the type of activity in each activity area, and the density of people on campus.

[0015] The campus intelligent management database module is used to determine the campus intelligent management database based on the security level of each activity area, the security matching coefficient of multiple adjacent activity areas, and deep learning of past campus security incidents.

[0016] The early warning module is used to determine the activity areas under early warning status based on the analysis of early warning signals in the campus intelligent management database.

[0017] The Collaborative Team module is used to determine collaborative teams based on the warning events in the activity area under warning status, the security level of surrounding activity areas, and the warning matching table.

[0018] In this embodiment of the invention, the method determines corresponding job identifiers based on the job information and corresponding personnel information of each campus personnel; determines the activity areas of multiple campus personnel based on each job identifier, the three-dimensional model of the campus, and time; determines the security level of each activity area based on its location, activity type, and the density of campus personnel; and determines the campus intelligent management database based on the security level of each activity area, the security matching coefficient of multiple adjacent activity areas, and deep learning of past campus security events. This approach ensures the accuracy of the campus intelligent management database by incorporating the security levels of each activity area, the security matching coefficient of multiple adjacent activity areas, and deep learning of past campus security events.

[0019] Therefore, in the campus intelligent management database, activity areas under warning status are determined based on the analysis of campus warning signals; collaborative teams are determined based on the warning events of activity areas under warning status, the security levels of surrounding activity areas, and the warning matching table. This achieves overall interaction between the warning events of activity areas under warning status, the security levels of surrounding activity areas, and the warning matching table, ensuring the accuracy of collaborative teams. Furthermore, based on the precise control of the warning event by the collaborative teams, the security levels of surrounding activity areas are fully considered, thereby improving the accuracy of intelligent early warning of campus security risks. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the intelligent early warning method for campus security risks in an embodiment of the present invention.

[0021] Figure 2 This is a flowchart illustrating step S11 of the intelligent early warning method for campus security risks in this embodiment of the invention.

[0022] Figure 3 This is a flowchart illustrating step S12 of the intelligent early warning method for campus security risks in this embodiment of the invention.

[0023] Figure 4 This is a flowchart illustrating step S13 of the intelligent early warning method for campus security risks in an embodiment of the present invention.

[0024] Figure 5 This is a flowchart illustrating step S14 of the intelligent early warning method for campus security risks in this embodiment of the invention.

[0025] Figure 6 This is a flowchart illustrating step S15 of the intelligent early warning method for campus security risks in an embodiment of the present invention.

[0026] Figure 7 This is a flowchart illustrating step S16 of the intelligent early warning method for campus security risks in an embodiment of the present invention.

[0027] Figure 8 This is a schematic diagram of the structural composition of the intelligent early warning system for campus security risks in an embodiment of the present invention. Detailed Implementation

[0028] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0029] Please see Figures 1 to 7 An intelligent early warning method for campus security risks is proposed and applied to intelligent early warning scenarios for campus security risks. The intelligent early warning method for campus security risks includes:

[0030] Step S11: Determine the corresponding job identifier based on the job information of each campus staff member and the corresponding personnel information;

[0031] Step S12: Determine the activity areas of multiple campus personnel based on the job identifiers, the 3D model of the campus, and the time.

[0032] Step S13: Determine the security level of each activity area based on its location, the type of activity, and the density of people on campus.

[0033] Step S14: Determine the campus intelligent management database based on the security level of each activity area, the security matching coefficient of multiple adjacent activity areas, and deep learning of past campus security incidents;

[0034] Step S15: In the campus intelligent management database, determine the activity areas under warning status based on the analysis of the campus's warning signals;

[0035] Step S16: Determine the collaborative team based on the warning events in the activity area under warning status, the security level of the surrounding activity areas, and the warning matching table.

[0036] refer to Figure 2 In step S11, the corresponding job identifier is determined based on the job information of each campus staff member and the corresponding personnel information.

[0037] In the specific implementation of this invention, the specific steps are as follows:

[0038] S111: Determine the personnel information of each campus staff member based on the campus roster and the personal database of campus personnel;

[0039] S112: Determine the job information of each campus staff member based on the campus staff database, the activity paths of each campus staff member, and the job responsibilities of each campus staff member;

[0040] S113: Determine the corresponding job identifier based on the job information of each campus staff member and the corresponding personnel information.

[0041] In the embodiments of this application, the personnel information of each campus person is determined based on the campus roster and the personal database of campus personnel, which takes into account the overall consideration of the campus roster and the personal database of campus personnel, and ensures the accuracy of the personnel information of each campus person.

[0042] At this time, a campus roster and a personal database of campus personnel were introduced. The campus roster is a list containing basic information about all teachers, students and staff of the school, including name, gender, age, class / department, student ID / employee ID, etc. The campus roster is compiled and maintained by the school's academic affairs department or human resources department.

[0043] The campus personnel personal database is a database system that stores detailed personal information of all faculty, staff, and students. This information includes contact information, home address, educational background, work experience, health status, etc. The information in the personal database comes from multiple channels such as new student registration, faculty and staff onboarding, and health check records.

[0044] Information matching and integration refers to combining basic information from the campus roster with detailed information from the personal database to form a complete and accurate campus personnel information database. At the same time, the basic information in the campus roster is matched with the detailed information in the personal database using unique identifiers such as student ID / employee ID. For data that cannot be matched, manual verification and supplementation are carried out. The integrated information database should contain comprehensive information on campus personnel to facilitate subsequent management and analysis.

[0045] Furthermore, the job information of each campus staff member is determined based on the campus staff database, their activity paths, and their job responsibilities. This comprehensive approach, which considers the campus staff database, their activity paths, and their job responsibilities, ensures the accuracy of the job information for each campus staff member.

[0046] At this point, basic information related to job information, such as name, employee number, department, and position, is extracted from the established campus personnel database. At the same time, a query statement is written using a database query language (such as SQL) to retrieve the required basic information from the database and export it as a temporary file or process it directly in memory.

[0047] By utilizing campus surveillance systems, access control systems, GPS positioning, and other technologies, we can track and analyze the activity paths of various campus personnel. We can collect surveillance footage, access control records, GPS positioning data, etc., over a period of time (such as one week or one month), and use data analysis software (such as Python, R, etc.) to process and analyze this data to extract information such as the daily activity areas and movement trajectories of campus personnel.

[0048] Based on the school's organizational structure, departmental responsibilities, and job descriptions, the job duties and requirements of each campus staff member are clearly defined. The extracted basic information, tracked activity paths, and clearly defined job responsibilities are combined to comprehensively determine the job information for each campus staff member. Optionally, a job information table is created, in which information such as the campus staff member's name, employee number, department, position, daily activity area, movement trajectory, and job responsibilities are entered. Simultaneously, the job information is further refined and categorized based on the job descriptions and the school's actual needs.

[0049] Therefore, by determining the corresponding job identifiers based on the job information and corresponding personnel information of each campus staff member, the interaction of job information and corresponding personnel information of each campus staff member is realized, and the job identifiers are further precisely controlled.

[0050] At this point, the job information (such as job title, job responsibilities, activity area, etc.) and personnel information (such as name, student ID / employee ID, contact information, etc.) are checked to remove duplicate or invalid information and supplement missing key information; considering the uniqueness, readability and memorability of job information, the identifier generation rules containing key job information are designed; for example, the format of "department abbreviation + job abbreviation + serial number" is adopted, or a part of the personnel information (such as the last few digits of the student ID / employee ID) is combined to generate the identifier.

[0051] Using built-in functions of programming languages ​​(such as Python, Java, etc.) or database management systems (such as MySQL, Oracle, etc.), job identifiers are automatically generated based on job information and personnel information; the generated job identifiers are verified to ensure their uniqueness and accuracy, and adjusted as needed.

[0052] Specifically, assuming a university needs to generate unique job identifiers for all faculty and students, the following are the specific steps:

[0053] The previous steps collected information on all faculty, staff, and students, including their names, student IDs / employee IDs, departments, and positions. For example, Professor Zhang from the Computer Science Department was identified as having employee ID 1234567 and a professor position.

[0054] Considering the school's actual needs and management requirements, the following generation rule is designed: The identifier is generated using the format "Department Abbreviation (2 digits) + Position Abbreviation (2 digits) + Last 4 digits of Student ID / Employee ID". For example, the Computer Science Department is abbreviated as "CS", and the professor's position is abbreviated as "PJ". Therefore, Professor Zhang's position identifier is generated as "CS + PJ + Last 4 digits of Employee ID". Based on the designed generation rule, a position identifier is generated for Professor Zhang. Professor Zhang's employee ID is 1234567, and the last 4 digits are 4567. Therefore, Professor Zhang's position identifier is "CSPJ4567".

[0055] Check for duplicates or errors in the generated job identifiers by querying the database or through manual verification. In this example, it is assumed that there are no duplicate employee numbers in the database, and the design of the department abbreviation and job title abbreviation is also unique. Therefore, the generated job identifier "CSPJ4567" is unique. If duplicates or errors are found in actual operation, the generation rules should be adjusted or the identifier should be regenerated in a timely manner.

[0056] refer to Figure 3 In step S12, the activity areas of multiple campus personnel are determined based on the job identifiers, the three-dimensional model of the campus, and the time.

[0057] In the specific implementation of this invention, the specific steps are as follows:

[0058] S121: Collect multiple images of the campus based on the dynamic shooting of the campus by drones, and determine the three-dimensional model of the campus based on the multiple images of the campus, the building distribution map of the campus, and the shape of the campus.

[0059] S122: Associate each job identifier with the campus 3D model and record the current position of each job identifier in the campus 3D model;

[0060] S123: Determine the first area parameters based on the current location of various job identifiers and the current work content corresponding to each job identifier; determine the second area parameters based on the current location and time of each job identifier; and determine the activity areas of multiple campus personnel based on the first area parameters, the second area parameters, and the three-dimensional model of the campus.

[0061] In the embodiments of this application, multiple images of the campus are collected based on dynamic photography of the campus by drones, and a three-dimensional model of the campus is determined based on the multiple images of the campus, the building distribution map of the campus, and the shape of the campus. This approach takes into account the overall consideration of the multiple images of the campus, the building distribution map of the campus, and the shape of the campus, thus ensuring the accuracy of the three-dimensional model of the campus.

[0062] At this point, the flight path of the drone is planned according to the campus area, terrain and building layout; the flight path is ensured to cover the entire campus, and the shooting needs at different heights and angles are taken into account; the camera parameters of the drone, such as exposure time, ISO, focal length, etc., are adjusted to ensure that the captured images are clear and the colors are accurate; at the same time, the shooting frequency and overlap are set for subsequent 3D reconstruction.

[0063] The drone flies along a planned flight path and captures images of the campus at preset positions and altitudes; ensuring sufficient image coverage for each area and appropriate overlap between images; the acquired images are preprocessed, including denoising, distortion correction, and contrast enhancement, to improve image quality, which is helpful for the subsequent 3D reconstruction process.

[0064] Obtain a building distribution map of the campus, including information such as the location, shape, and height of buildings, which serves as a reference for 3D reconstruction. In addition to the building distribution map, other morphological features of the campus, such as topographic relief and vegetation distribution, need to be considered. This information is obtained through ground images taken by drones or on-site surveys. Using 3D reconstruction techniques such as Structure from Motion (SFM) or Multi-View Stereo Vision (MVS), combined with the preprocessed images, building distribution map, and campus morphological information, a 3D model of the campus is generated. The generated 3D model is then optimized, including smoothing surfaces, filling holes, and adjusting scale, to ensure the accuracy and realism of the model.

[0065] Furthermore, each job identifier is associated with a 3D model of the campus, and the current position of each job identifier is recorded in the 3D model. Based on the current position of each job identifier and the current work content corresponding to each job identifier, the first area parameter is determined. Based on the current position and time of each job identifier, the second area parameter is determined. Based on the first area parameter, the second area parameter, and the 3D model of the campus, the activity areas of multiple campus personnel are determined. This method takes into account the overall consideration of the first area parameter, the second area parameter, and the 3D model of the campus, ensuring the accuracy of the activity areas of multiple campus personnel.

[0066] At this point, it is necessary to determine how to associate job identifiers with specific locations in the campus 3D model. This involves matching job identifiers with specific areas within the campus (such as buildings, floors, rooms, etc.). Based on the determined association rules, each job identifier is associated with its corresponding location in the campus 3D model. This is achieved by marking the specific location of the job on the 3D model, for example, using coordinate points, polygonal regions, or other identifiers.

[0067] Choose an appropriate positioning technology to obtain the current location of each job identifier. This could be GPS-based positioning, Wi-Fi-based positioning, Bluetooth-based positioning, or a hybrid positioning method that combines multiple technologies.

[0068] Using the selected positioning technology, the current location data of each job identifier is collected in real time. This involves interfaceing with the positioning system to obtain real-time location information. The collected current location data of the job identifiers is then mapped onto the campus 3D model. This involves converting the location data (such as latitude and longitude, coordinate points, etc.) into the corresponding positions in the 3D model. Based on the mapped location data, the campus 3D model is updated to reflect the current position of each job identifier. This is achieved by adding or updating markers, icons, or other visual elements on the 3D model.

[0069] Furthermore, a first area parameter, a second area parameter, and a three-dimensional model of the campus were introduced. For the first area parameter, the current location of each job identifier and its corresponding current work content were collected. This involves interface docking with the campus management system or personnel database to obtain real-time job information and work content.

[0070] The relationship between job content and the current location of job identifiers is analyzed by statistically analyzing the working time, frequency, or type of each job within a specific area. Based on the analysis results, the first-region parameters closely related to the job content are determined. The first-region parameters include the work density, job type distribution, and working time period within a specific area.

[0071] For the second region parameters, the current location and time data of each job identifier are collected, which is achieved through positioning technology (such as GPS, Wi-Fi positioning, etc.) and timestamp recording; the relationship between the location data and time data of the job identifiers is analyzed, which reveals the movement patterns, dwell time or activity hotspots of personnel in different time periods; based on the analysis results, the second region parameters closely related to time are determined; the second region parameters include activity hotspots, personnel flow patterns, and dwell time distribution in specific time periods.

[0072] The parameters of the first and second regions are integrated to form a comprehensive set of region parameters. This involves taking into account information from multiple aspects such as work content, time, and location. The integrated region parameters are then applied to a three-dimensional model of the campus. This is achieved by marking specific areas on the model, using color coding, or adding annotations. Based on the markings and annotations on the three-dimensional model, multiple activity areas for campus personnel are determined. These areas are specific buildings, floors, rooms, or outdoor areas, and are further subdivided and named as needed.

[0073] refer to Figure 4 In step S13, the security level of each activity area is determined based on the location of each activity area, the type of activity in each activity area, and the density of people on campus.

[0074] In the specific implementation of this invention, the specific steps are as follows:

[0075] S131: Mark the location of each activity area and determine multiple activities in each activity area based on the tracing of the location of each activity area.

[0076] S132: In each activity area, the activity type is determined by deep learning based on multiple activity contents, corresponding campus personnel, and corresponding activity scope;

[0077] S133: Determine the density of campus personnel based on the area of ​​each activity zone, the distribution of corresponding campus personnel, and the number of campus personnel.

[0078] S134: Determine the security level of each activity area based on its location, the type of activity, and the density of people on campus.

[0079] In the embodiments of this application, the regional location of each activity area is marked, and multiple activity contents in each activity area are determined by tracing the regional location of each activity area, thereby ensuring the accuracy of multiple activity contents in each activity area.

[0080] At this stage, basic information about each activity area on campus is collected, including area name, boundary range, and geographical location. This information is obtained through campus maps, architectural drawings, or on-site surveys. Based on the collected area information, the location of each activity area is accurately marked on the campus's 3D model or floor plan. This involves drawing boundary lines on the model or drawings, adding labels, or using color coding to represent different areas. After marking is completed, it needs to be verified to ensure the accuracy of the marking. This is achieved by confirming with campus administrators, students, or faculty members, or by using positioning technology (such as GPS) for on-site measurement.

[0081] Historical data from various activity areas is collected and analyzed, including past activity records, personnel flow, and equipment usage. This data comes from the campus management system, surveillance cameras, access control systems, etc. By analyzing historical data, the various activities occurring in each activity area are identified. For example, in the teaching building, activities include lectures, seminars, student self-study, and exams; in the library, activities include reading, borrowing books, and attending lectures or seminars. Further analysis of historical data determines the type and frequency of each activity in a specific area, which helps to understand the main functions and usage of each activity area.

[0082] Furthermore, within each activity area, the activity type is determined based on deep learning of multiple activity contents, corresponding campus personnel, and corresponding activity scope. This approach incorporates a holistic consideration of deep learning of multiple activity contents, corresponding campus personnel, and corresponding activity scope, ensuring the accuracy of activity types in each activity area.

[0083] At this point, records of activities taking place in each activity area are collected. These records come from the campus management system, timetables, activity schedules, surveillance videos, etc. Information on campus personnel participating in these activities is collected, including their identities (such as students, teachers, visitors, etc.), professional backgrounds, and level of participation. The specific scope of each activity is defined, which is obtained through campus maps, activity site layout maps, etc.

[0084] The collected activity data, personnel information, and activity scope data are preprocessed, including data cleaning, format standardization, and feature extraction. A suitable deep learning model, such as a convolutional neural network (CNN), a recurrent neural network (RNN), or their variants, is selected to analyze the activity data and determine the activity type. The deep learning model is trained using the preprocessed data. During the training process, the model will learn how to extract key information from features such as activity content, participants, and activity scope, and determine the activity type accordingly.

[0085] When a new activity occurs, collect relevant data on the activity content, participants, and scope; input the new data into a trained deep learning model, which will output a predicted activity type; validate the model's predictions to ensure their accuracy; and, if necessary, adjust and optimize the model based on the validation results.

[0086] Furthermore, the density of campus personnel is determined based on the area of ​​each activity zone, the distribution location of the corresponding campus personnel, and the number of campus personnel. This takes into account the overall consideration of the area of ​​each activity zone, the distribution location of the corresponding campus personnel, and the number of campus personnel, ensuring the accuracy of the density of campus personnel.

[0087] At this point, the area of ​​each activity area is accurately measured, which is obtained through campus maps, architectural drawings, or on-site measurements; real-time distribution and number of campus personnel within each activity area are collected, and this information comes from campus positioning systems (such as Wi-Fi positioning, Bluetooth positioning, or GPS positioning), access control systems, student / faculty databases, etc.

[0088] The collected data on area and personnel distribution and quantity were preprocessed to ensure accuracy and consistency. The formula "Personnel Density = Number of Personnel / Area" was used to calculate the personnel density in each activity area. This formula provides a metric for assessing the density of personnel in each activity area. The changing trends of personnel density in each activity area over different time periods were analyzed to understand the patterns and characteristics of personnel movement on campus. Based on the analysis results of personnel density, the safety risks and emergency needs of each activity area were assessed to provide decision support for campus management.

[0089] Therefore, the safety level of each activity area is determined based on its location, the type of activity, and the density of people on campus. This allows for interaction between the location, activity type, and density of people on campus, enabling more precise control over the safety level of each activity area.

[0090] At this point, collect specific location information of each activity area within the campus, including their relative positions, distances from campus entrances and exits, and the complexity of the surrounding environment; based on previous analysis, identify the main types of activities in each activity area, such as academic activities, sports activities, and catering services; and based on real-time or historical data, understand the density of campus personnel in each activity area, including peak hours and normal personnel distribution.

[0091] Analyze the impact of the location of each activity area on its safety; for example, areas near campus entrances or in complex surrounding environments face a higher risk of external threats; assess the safety risks associated with different types of activities; for example, sports activity areas pose a risk of injury due to the use of sports equipment and personnel activities; academic activity areas pose a risk of evacuation difficulties due to large gatherings of people; consider the impact of population density on safety risks; high-density areas are more difficult to evacuate and rescue quickly and effectively in emergencies.

[0092] A comprehensive assessment of location risk, activity type risk, and population density risk should be conducted, taking into account their interactions. For example, an area located near an entrance or exit but with relatively safe activities and moderate population density has a lower security risk than an area located in a secluded location but with high-risk activities and high population density.

[0093] Based on the comprehensive assessment results, each activity area is divided into different safety levels, such as "high risk", "medium risk" and "low risk". These levels are used to guide the formulation and implementation of campus safety management measures.

[0094] In one embodiment of this application, a preset security level matching table is collected, as shown in Table 1:

[0095] Table 1 Security Level Matching Table

[0096] ,

[0097] In this example, teaching building L has a low safety risk under normal circumstances because it is located in the center and mainly used for academic activities. However, during peak hours, due to the high density of people, its safety level is medium risk. Gymnasium M has a high safety level because it is located on the periphery and mainly used for sports activities (with a risk of sports injuries). In addition, the density of people during activities is moderate.

[0098] Library N is located in a quiet area and is primarily used for reading activities, so its safety risk is low under normal circumstances. However, the density of people increases during peak hours, and therefore the safety level will be raised accordingly. Cafeteria O has an extremely high safety level due to the extremely high density of people during mealtimes.

[0099] refer to Figure 5 In step S14, the campus intelligent management database is determined based on the security level of each activity area, the security matching coefficient of multiple adjacent activity areas, and deep learning of past security incidents on campus.

[0100] In the specific implementation of this invention, the specific steps are as follows:

[0101] S141: Determine the activity area to be managed based on the security level of each activity area and the preset security level threshold, and determine the security matching coefficient of the multiple adjacent activity areas based on the matching of the activity area to be managed and the corresponding multiple adjacent activity areas.

[0102] S142: Collect a 3D model of the campus and determine past security incidents of the campus based on the 3D model of the campus, the activity areas to be managed, and the corresponding multiple adjacent activity areas.

[0103] S143: Deep learning is performed on the security levels of each activity area, the security matching coefficients of multiple adjacent activity areas, and past security incidents on campus. Based on the deep learning of the security levels of each activity area, the security matching coefficients of multiple adjacent activity areas, and past security incidents on campus, a campus intelligent management database is determined.

[0104] In the embodiments of this application, the activity areas to be managed are determined according to the security level of each activity area and the preset security level threshold, and the security matching coefficient of the adjacent activity areas is determined based on the matching of the activity areas to be managed and the corresponding adjacent activity areas. This takes into account the overall consideration of the matching of the activity areas to be managed and the corresponding adjacent activity areas, and ensures the accuracy of the security matching coefficient of the adjacent activity areas.

[0105] At this point, security level information for each activity area is collected. This information is calculated through the previous steps (such as S134). Based on the needs of campus security management, a series of security level thresholds are set. These thresholds are specific values ​​or level labels used to determine which activity areas have higher security risks and require special attention or control. The security level of each activity area is compared with the preset security level thresholds. If the security level of an activity area exceeds or equals a certain threshold, it is identified as an activity area to be controlled.

[0106] For each activity area to be managed, it is necessary to identify multiple adjacent activity areas. These adjacent areas are physically adjacent and also functionally or procedurally related. Based on the security correlation between the adjacent areas and the area to be managed, a security matching coefficient is calculated between them. This coefficient is a value between 0 and 1, representing the degree of security matching or mutual influence between the adjacent areas and the area to be managed.

[0107] When calculating the safety matching coefficient, multiple factors are considered, such as the types of activities in adjacent areas, the flow of people, and the effectiveness of safety management measures. These factors are quantified through historical data, expert evaluation, or machine learning.

[0108] Specifically, suppose there are five activity areas on a university campus: teaching building A, library B, gymnasium C, canteen D, and dormitory E; teaching building A is classified as "high risk", library B as "medium risk", gymnasium C as "low risk", canteen D as "medium risk", and dormitory E as "low risk"; "medium risk" and above are set as the safety level threshold requiring special attention; based on the comparison between the safety level and the threshold, teaching building A, library B, and canteen D are identified as activity areas to be controlled.

[0109] The adjacent areas of teaching building A are library B, gymnasium C, and canteen D; the adjacent areas of library B are teaching building A, gymnasium C, and dormitory E; the safety matching coefficient between teaching building A and library B is 0.8 because both are located in the central area of ​​the campus, with frequent personnel flow and both involve academic activities, resulting in a high degree of safety correlation; the safety matching coefficient between teaching building A and gymnasium C is 0.5 because although gymnasium C is adjacent to teaching building A, it is mainly used for sports activities and has a lower correlation with the academic activities of the teaching building.

[0110] The safety matching coefficient between Teaching Building A and Cafeteria D is 0.6, because Cafeteria D is the main place for teachers and students to eat, and there is some overlap with Teaching Building A in terms of personnel flow and time. The safety matching coefficient between Library B and Gymnasium C is 0.4, because sports activities in Gymnasium C have a certain impact on the reading environment of the library, but the overall correlation is not high. The safety matching coefficient between Library B and Dormitory E is 0.7, because Dormitory E is the main living area for students, and there is some connection with Library B in terms of academic activities and personnel flow.

[0111] Furthermore, a 3D model of the campus is collected, and based on the 3D model, the activity areas to be managed, and the corresponding adjacent activity areas, past security incidents on the campus are determined. This enables interaction between the 3D model of the campus, the activity areas to be managed, and the corresponding adjacent activity areas, allowing for more precise control over past security incidents on the campus.

[0112] At this point, modern surveying technologies (such as drone aerial photography, LiDAR scanning, and 3D modeling software) are used to collect a three-dimensional model of the campus. These technologies can capture the accurate location and shape information of elements such as buildings, roads, and green belts within the campus. During the data collection process, it is necessary to ensure the integrity and accuracy of the data, including information such as the height, shape, and material of buildings, the width and direction of roads, and the distribution of green belts. The collected data is then imported into 3D modeling software to construct a three-dimensional model of the campus. This model should be able to realistically reflect the spatial structure and layout of the campus.

[0113] Review the historical records of the campus security management department to understand the circumstances of various security incidents (such as fires, thefts, and fights) that have occurred on campus in the past; use a three-dimensional model of the campus to locate historical security incidents to specific activity areas, which is achieved by matching the descriptions of the incidents with elements in the model; for each located security incident, analyze its causes, processes, impacts, and corresponding countermeasures, which helps to understand the distribution patterns and influencing factors of security incidents on campus; consider the interactions between the activity areas to be managed and their adjacent areas; analyze the connections between these areas in terms of function, personnel flow, and time, as well as their impact on security incidents.

[0114] Therefore, deep learning is used to study the security levels of each activity area, the security matching coefficients of multiple adjacent activity areas, and past security incidents on campus. Based on the deep learning of the security levels of each activity area, the security matching coefficients of multiple adjacent activity areas, and past security incidents on campus, the campus intelligent management database is determined. This database is compatible with the deep learning of the security levels of each activity area, the security matching coefficients of multiple adjacent activity areas, and past security incidents on campus, thus ensuring the accuracy of the campus intelligent management database.

[0115] At this point, the security levels of each activity area, the security matching coefficients of adjacent activity areas, and data on past security incidents on campus are collected. This data comes from different data sources, such as the campus security management system, student information system, and faculty and staff information system. The collected data is cleaned to remove duplicate, erroneous, or invalid data. The accuracy and consistency of the data are ensured. The data is then converted into a format suitable for deep learning models to process, which includes converting categorical data into numerical data and standardizing or normalizing the data.

[0116] Choose an appropriate deep learning model architecture based on the complexity of the problem and the characteristics of the data; for example, use convolutional neural networks (CNNs) to process image data, recurrent neural networks (RNNs) or long short-term memory networks (LSTMs) to process time series data, or multilayer perceptrons (MLPs) to process general numerical data; choose an appropriate loss function based on the type of problem (e.g., classification, regression); simultaneously, choose an appropriate optimizer to minimize the loss function, such as stochastic gradient descent (SGD) or Adam; input the preprocessed data into the deep learning model for training; during training, the model continuously adjusts its parameters to minimize the loss function and improve prediction accuracy.

[0117] Cross-validation is used to evaluate model performance. This involves splitting the data into training and test sets, training the model on the training set, and evaluating its performance on the test set. Multiple cross-validations yield more reliable performance evaluations. Model performance is further optimized by adjusting hyperparameters (such as learning rate, batch size, and number of network layers), achieved through methods like grid search, random search, or Bayesian optimization. For complex deep learning models, model interpretation techniques are needed to understand their decision-making processes, which helps increase the model's transparency and credibility.

[0118] Using a trained deep learning model, the security levels of various activity areas, the security matching coefficients of adjacent activity areas, and past security incidents on campus are predicted. These predictions include the probability distribution of security risks and potential security vulnerabilities. The prediction results are then integrated with other relevant information (such as the geographical location of activity areas, personnel flow, and security management measures) to construct a campus intelligent management database. This database provides a comprehensive perspective for understanding and managing the campus security situation. Based on the information in the intelligent management database, targeted security management strategies are formulated, including strengthening security monitoring in certain areas, optimizing personnel flow paths, and improving security awareness and training.

[0119] Specifically, suppose a university campus has five activity areas: teaching building A, library B, gymnasium C, canteen D, and dormitory E; collect the security level (e.g., high risk, medium risk, low risk) of each activity area, the security matching coefficient of adjacent activity areas (e.g., 0.8, 0.5, etc.), and data on past campus security incidents (e.g., fires, thefts, etc.); clean and format the data to ensure its accuracy and consistency; for example, convert the security level into numerical data (high risk = 3, medium risk = 2, low risk = 1), and retain the security matching coefficient in decimal form.

[0120] Multilayer Perceptron (MLP) was chosen as the deep learning model architecture because it can handle general numerical data; a cross-entropy loss function and Adam optimizer were defined to train the model; preprocessed data was input into the MLP model for training; 5-fold cross-validation was used to evaluate the model's performance; through multiple cross-validations, the model's average accuracy reached over 90%; hyperparameters of the model were tuned, such as adjusting the learning rate and batch size, to further improve the model's performance; model interpretation techniques were used to understand the model's decision-making process, revealing that the model mainly relies on safety level and safety matching coefficient for prediction.

[0121] Using a trained MLP model, the safety levels of each activity area, the safety matching coefficients of adjacent activity areas, and past campus safety incidents are predicted. The predicted risks are: teaching building A has a high safety risk; library B and gymnasium C have a medium safety risk; and canteen D and dormitory E have a low safety risk. These predictions are then integrated with other relevant information to construct a campus intelligent management database. This database contains information such as the safety risk level of each activity area, potential safety hazards, and safety management measures. Based on the information in the intelligent management database, targeted safety management strategies are developed. For example, security monitoring of teaching building A is strengthened, personnel flow paths are optimized to reduce congestion and conflict, and safety awareness and training for all teachers and students are improved.

[0122] refer to Figure 6 In step S15, the activity areas under warning status are determined in the campus intelligent management database based on the analysis of the campus's warning signals.

[0123] In the specific implementation of this invention, the specific steps are as follows:

[0124] S151: Real-time monitoring of the campus intelligent management database;

[0125] S152: In the campus intelligent management database, output the corresponding early warning signal based on the campus intelligent management database, and determine the corresponding early warning information based on the tracing of the early warning signal;

[0126] S153: Based on the identification of early warning information, multiple activity areas are determined and synchronous monitoring of multiple activity areas is triggered;

[0127] S154: Output multiple activity warning features based on the synchronous monitoring of multiple activity areas, and determine the activity areas in the warning state based on the location of the multiple activity warning features, the influence range of the multiple activity warning features, and the spatial location of multiple campus personnel.

[0128] In the embodiments of this application, the campus intelligent management database is monitored in real time; based on the campus intelligent management database, corresponding early warning signals are output, and the corresponding early warning information is determined by tracing the early warning signals, thus ensuring the accuracy of the early warning information.

[0129] At this point, the campus intelligent management database is monitored in real time, and subsequent control measures are implemented. A series of early warning rules are pre-set in the campus intelligent management database. These rules are formulated based on historical data, safety standards, regulatory requirements, and other factors to identify potential security risks or anomalies. The early warning rules involve multiple data dimensions, such as the security level of the activity area, environmental monitoring data (e.g., temperature, smoke concentration), personnel flow, and access control system records.

[0130] The system collects and processes information from various data sources within the campus in real time, including sensor data, video surveillance data, and personnel location data. It uses data analysis techniques (such as machine learning and statistical models) to process and analyze the collected data in order to identify whether the conditions set in the early warning rules are met.

[0131] When the system detects that the data meets the warning rules, it automatically outputs a warning signal. The warning signal can take many forms, such as sound alarms, visual cues (e.g., flashing red lights), text messages, email notifications, etc. The warning signal should contain enough information so that managers can quickly identify the type and urgency of the warning.

[0132] Once an early warning signal is triggered, the system should be able to automatically trace and determine the corresponding early warning information, including the time and location of the warning, the area of ​​activity involved, the cause, and the suggested response measures. The system should record detailed information for each early warning, including the warning signal, warning information, response measures, and handling results. Early warning reports should be generated regularly to summarize and analyze early warning events so that management personnel can understand the safety status on campus and implement improvement measures.

[0133] Specifically, suppose the following early warning rules are set in the intelligent management database of a university campus:

[0134] A fire alarm is triggered when the smoke concentration in an activity area exceeds a certain threshold; a crowding alarm is triggered when the population density in an area exceeds a set value; and an intrusion alarm is triggered when the access control system records that an important area has been illegally entered outside of working hours.

[0135] One evening, the campus intelligent management system detected a sudden increase in smoke concentration in the library area, exceeding the preset threshold. The system immediately triggered a fire alarm signal, which was manifested by a flashing red light on the monitoring interface and an SMS notification to campus security personnel. After receiving the alarm signal, the personnel immediately logged into the campus intelligent management database to view the alarm information.

[0136] The warning information shows:

[0137] Warning type: Fire warning; Time of occurrence: 8:30 PM; Location: Library area; Cause: Smoke concentration exceeds the standard (specific value); Recommended response: Immediately evacuate people from the library, activate the fire protection system, and contact the fire department.

[0138] Based on the early warning information, the management personnel acted swiftly, evacuating the library staff and activating the fire suppression system. They also contacted the fire department for assistance. An inspection revealed a small fire caused by aging wiring in the library; thanks to the timely detection and handling, no serious consequences resulted. Following this incident, campus security personnel conducted a safety inspection of the library area and rectified issues such as aging wiring to prevent similar incidents from recurring. Simultaneously, the campus intelligent management database recorded detailed information about the incident and generated an early warning report for subsequent analysis and improvement by management.

[0139] Furthermore, multiple activity areas are identified based on the early warning information, and synchronous monitoring of these multiple activity areas is triggered, ensuring the accuracy of the multiple activity areas.

[0140] At this point, when the campus intelligent management database issues an early warning signal, the first step is to analyze the early warning information in detail. This includes identifying the type of early warning (such as fire, theft, overcrowding, etc.), the time and location of the incident, and the multiple activity areas involved. When analyzing the early warning information, it is necessary to use the spatial data and logical relationships in the database to determine the scope of the impact of the early warning event, thereby identifying all relevant activity areas.

[0141] Based on the analysis results of the early warning information, multiple activity areas involved in the early warning are identified. These areas include the location where the early warning occurred, as well as adjacent or affected areas. When identifying activity areas, factors such as the function of the area, population density, and security level need to be considered to ensure the targeting and effectiveness of the monitoring.

[0142] Once multiple activity areas involved in the warning have been identified, simultaneous monitoring of these areas should be triggered immediately. Simultaneous monitoring includes various methods such as video surveillance, audio surveillance, environmental monitoring, and personnel positioning. When triggering simultaneous monitoring, it is necessary to ensure the stability and reliability of the monitoring system so that it can quickly provide clear and accurate information in emergency situations. In the process of triggering simultaneous monitoring, it is necessary to integrate various monitoring resources within the campus, including video cameras, audio acquisition devices, and environmental monitoring sensors in different locations.

[0143] When integrating monitoring resources, it is necessary to consider the optimal allocation and collaborative work of resources to ensure the comprehensiveness and efficiency of monitoring.

[0144] During synchronous monitoring, monitoring information is recorded in real time, including video recordings, audio recordings, and environmental monitoring data; the monitoring information is analyzed and processed to extract key clues, assess the security situation, and provide support for subsequent response and decision-making.

[0145] Specifically, suppose a university campus intelligent management database issues a fire warning signal. The warning information indicates that the fire occurred in the library area and affected the adjacent study room area and canteen area. After receiving the warning signal, the management personnel first analyze the warning information to confirm that the fire occurred in the library area and identify the affected study room area and canteen area. Based on the analysis results of the warning information, the activity areas involved in the warning are determined to be the library area, study room area and canteen area.

[0146] The system immediately triggered simultaneous monitoring of the three areas. In the library area, video surveillance and smoke concentration monitoring were activated. In the study room and cafeteria areas, video surveillance and personnel location monitoring were activated. The system integrated video surveillance cameras, smoke concentration sensors, and personnel location devices in the library, study room, and cafeteria areas to ensure comprehensive and efficient monitoring. During the simultaneous monitoring, video recordings, smoke concentration data, and personnel location information were recorded in real time. By analyzing the monitoring information, the management personnel discovered that the fire was spreading in the library area, some people had been evacuated from the study room area, but a large number of people were still gathered in the cafeteria area.

[0147] Therefore, multiple activity warning features are output based on the synchronous monitoring of multiple activity areas, and the activity areas in the warning state are determined based on the location of the multiple activity warning features, the influence range of the multiple activity warning features, and the spatial location of multiple campus personnel.

[0148] At this point, after simultaneous monitoring is initiated in multiple activity areas, the system begins to collect and analyze monitoring data from these areas. This includes video surveillance recordings, audio recordings, environmental monitoring data (such as smoke concentration and temperature), and personnel location information. When analyzing the monitoring data, the system uses technologies such as image recognition, sound analysis, and data statistics to extract feature information related to the warning event. Based on the analysis results of the simultaneous monitoring data, the system extracts multiple activity warning features, including abnormal behavior patterns (such as running and screaming), exceeding environmental monitoring data standards (such as a sharp increase in smoke concentration), and abnormal gathering or evacuation of personnel. Each warning feature contains key elements such as specific location information, timestamp, and scope of impact.

[0149] For each extracted warning feature, the system needs to determine its specific location and scope of influence. This is achieved by analyzing spatial information in the monitoring data, such as the location of objects in the video and the trajectory of people's movement. By using the campus personnel positioning system or other relevant technologies, the system can determine the spatial location of people on campus in real time. This helps managers understand which personnel are in danger zones or need to be evacuated when a warning event occurs. Combining the above information, the system can determine which activity areas are in a warning state. This includes the activity area where the warning feature is located, as well as other adjacent or related areas affected by the warning feature. Activity areas in a warning state require special attention, and corresponding safety management measures should be taken, such as evacuating personnel and activating emergency response procedures.

[0150] Specifically, suppose a university's campus intelligent management system triggers a fire alarm in the library area and initiates simultaneous monitoring of the adjacent study room and cafeteria areas. The system begins collecting and analyzing monitoring data from the library, study rooms, and cafeteria areas. In the library area, video surveillance shows smoke filling the area, and people begin to evacuate. In the study room area, location information shows a large number of students moving towards the exits. In the cafeteria area, although no obvious signs of fire have been found, people are beginning to show signs of unease. The system extracts the following warning characteristics: smoke concentration exceeding the standard in the library area (specific values), abnormal evacuation of people in the study room area (number of people and speed), and slight disturbances among people in the cafeteria area (derived through audio analysis).

[0151] The smoke concentration exceeding the standard is located in the central part of the library area, and the affected area is spreading outwards. Abnormal evacuation is mainly occurring in the study room area near the library, affecting the entire study room area. Minor disturbances are present in the cafeteria area, but no clear evacuation trend has yet emerged. The personnel positioning system has identified XX students and XX faculty / staff members in the library area, XX students in the study room area, and XX diners in the cafeteria area; their specific locations are also being recorded in real time. Based on the above information, the system has determined that the library area and study room area are under alert. Due to the excessive smoke concentration and personnel evacuation in the library area, fire extinguishing and evacuation procedures need to be initiated immediately. The study room area, being adjacent to the library and with a large number of people already evacuated, also needs to initiate evacuation procedures and prepare for emergencies. Although no obvious signs of fire have been found in the cafeteria area, considering the growing anxiety among the occupants, evacuation preparations are also recommended.

[0152] In one embodiment of this application, a preset warning feature matching table is collected, as shown in Table 2:

[0153] Table 2 Early Warning Feature Matching Table

[0154] ,

[0155] Suppose that monitoring data shows: smoke concentration in the central part of the library exceeds the standard; students in the study room area are moving rapidly towards the exit, indicating abnormal evacuation; video surveillance in the cafeteria kitchen area captures flames; the library is identified as being in an alert state because the smoke concentration exceeds the standard; the study room is also identified as being in an alert state because the abnormal evacuation of personnel matches the characteristic; and parts of the cafeteria are identified as being in an alert state because the abnormal (fire) characteristic matches the video surveillance.

[0156] refer to Figure 7 In step S16, the collaborative team is determined based on the warning events of the activity area in the warning state, the security level of the surrounding activity areas, and the warning matching table;

[0157] In the specific implementation of this invention, the specific steps are as follows:

[0158] S161: Output multiple sub-activity events based on monitoring of the activity area in the early warning state;

[0159] S162: Based on multiple sub-events, the degree of injury to campus personnel, and the number of personnel in each campus, determine the warning events for the activity area under warning status, determine the corresponding surrounding activity areas based on the location of the activity area under warning status, and determine the corresponding safety level by conducting safety inspections on the surrounding activity areas.

[0160] S163: Collect the early warning matching table and associate the early warning events of the activity area under early warning status and the security level of the surrounding activity areas with the early warning matching table;

[0161] S164: Determine the collaborating personnel for the surrounding activity areas based on the warning events of the activity area under warning status and the security level of the surrounding activity areas; determine the person in charge of the activity area under warning status based on the warning events of the activity area under warning status and the warning matching table.

[0162] S165: Determine the collaborative team based on the person in charge of the activity area under warning, the collaborators in the surrounding activity areas, and the campus personnel database.

[0163] In the embodiments of this application, multiple sub-activity events are output based on the monitoring of the activity area in the early warning state, thus introducing multiple sub-activity events.

[0164] When an activity area is identified as being under warning, the system first collects monitoring data for that area, including video surveillance recordings, audio recordings, environmental monitoring data (such as smoke concentration and temperature), and personnel location information. Based on the analysis of the monitoring data, the system identifies and extracts multiple sub-activity events directly related to the warning event. These sub-activity events are specific manifestations or components of the warning event. The identification of sub-activity events involves various technical means such as image recognition, sound analysis, and data statistics.

[0165] The system categorizes and labels the identified sub-events, which helps managers quickly understand the full picture of the warning event and take appropriate countermeasures. The categorization is based on factors such as the nature of the sub-event (e.g., fire, theft, personnel evacuation) and the scope of impact (e.g., local, regional, school-wide). The system outputs the identified and categorized sub-events to managers or relevant emergency response teams through displays, alarm systems, SMS, or email. At the same time, the system records these sub-events and their related information in a database for subsequent analysis and traceability.

[0166] Furthermore, based on multiple sub-events, the degree of injury to campus personnel, and the number of personnel in each campus, the warning events for the activity areas under warning status are determined. Based on the location of the activity areas under warning status, the corresponding surrounding activity areas are determined, and the corresponding safety level is determined by safety inspection of the surrounding activity areas, thus ensuring the accuracy of the safety level.

[0167] At this point, a comprehensive analysis of multiple sub-events is conducted to clarify the specific type and nature of the warning event, which involves various situations such as fire, explosion, theft, and personnel conflict. The system combines the degree of injury and the number of injured personnel on campus to further assess the severity of the warning event, which helps to determine the priority of emergency response and the resources required. Based on the analysis of sub-events and the assessment of personnel injury, the system comprehensively judges the warning events in the activity area under warning status and provides a clear description of the warning event.

[0168] The system uses a geographic information system (GIS) or a campus layout map to identify the geographical relationship between the activity area under warning and its surrounding activity areas; based on the geographical relationship, the system determines the specific scope of the surrounding activity areas, which includes adjacent classrooms, laboratories, dormitories, etc.

[0169] The system analyzes the potential risks that early warning events pose to surrounding activity areas, including fire spread, smoke diffusion, and public panic. Based on the analysis of potential risks, the system assesses the safety level of surrounding activity areas, which involves consideration of multiple dimensions, such as population density, building structure, and evacuation routes. The system sets clear safety level standards, such as "safe," "low risk," "medium risk," and "high risk," so that managers can quickly understand the safety status of the surrounding areas.

[0170] Specifically, suppose a university campus intelligent management system triggers a fire warning in the library area and identifies the library as an activity area under warning. The system analyzes multiple sub-events in the library area, including an open flame near a bookshelf on the second floor and abnormal evacuation of people in the library. Combining on-site surveillance video and injury reports, the system assesses the degree and number of injuries, finding that two students were injured during the evacuation, but their injuries were not serious. Based on the above information, the system determines that the warning event is "a small-scale fire occurred on the second floor of the library, accompanied by minor casualties."

[0171] The system used a campus layout map to identify activity areas around the library, including adjacent study rooms, teaching buildings, and dormitories; it also determined the specific boundaries of these surrounding areas for subsequent security level assessments.

[0172] The system analyzed the potential risks posed by the fire to the surrounding area, such as smoke spreading to study rooms and teaching buildings, and panic leading to difficulties in evacuation. Based on the analysis of potential risks, the system conducted a safety level assessment of the surrounding area. The study rooms, due to their close connection with the library and high population density, were assessed as "medium risk." The teaching buildings and dormitories, due to their greater distance, good building structure, and unobstructed evacuation routes, were assessed as "low risk." The system set clear safety level standards and notified management personnel and relevant emergency response teams via display screens or SMS messages so that they could take appropriate safety measures.

[0173] Furthermore, an early warning matching table is collected, and the early warning events of the activity area under early warning status and the security levels of the surrounding activity areas are associated with the early warning matching table. Based on the early warning events of the activity area under early warning status and the security levels of the surrounding activity areas, the collaborating personnel of the surrounding activity areas are determined, and based on the early warning events of the activity area under early warning status and the early warning matching table, the person in charge of the activity area under early warning status is determined, thus introducing the collaborating personnel of the surrounding activity areas and the person in charge of the activity area under early warning status.

[0174] At this point, the warning events of the activity area under warning status and the security levels of surrounding activity areas are associated with the warning matching table, thus establishing a connection between the three. Simultaneously, the collaborating personnel in the surrounding activity areas are determined based on the warning events of the activity area under warning status and the security levels of surrounding activity areas.

[0175] The system analyzes the type and number of personnel needed for coordination based on the security level of the surrounding activity area. The higher the security level, the more personnel are needed, and the more professional rescue or evacuation teams are required. The system matches suitable personnel with available resources on campus, such as security personnel, firefighters, and medical rescue teams, taking into account factors such as personnel's geographical location, professional skills, and availability. The system sends task assignment instructions to the selected personnel, specifying their responsibilities, routes of action, assembly points, and other key information to ensure they can respond quickly and effectively to early warning events.

[0176] Based on the warning events in the activity areas under warning status and the warning matching table, the system determines the person in charge of the activity areas under warning status. At this time, the system will search for the person in charge information corresponding to the warning event type and the warning matching table. The warning matching table contains the correspondence between different warning event types and persons in charge, as well as the person in charge's contact information and scope of responsibilities. After selecting the person in charge, the system will verify their current status, including whether they are online, whether they are nearby, and whether they have other urgent tasks. This helps ensure that the selected person in charge can respond immediately and effectively command. If the person in charge is in good condition and can respond, the system will send them detailed warning event information and instructions, including the event type, location, scope of impact, and information on collaborating personnel, and authorize them to be responsible for commanding and coordinating on-site emergency response work.

[0177] Therefore, the collaborative team is determined based on the person in charge of the activity area under warning, the coordinators in the surrounding activity areas, and the campus personnel database. This takes into account the overall consideration of the person in charge of the activity area under warning, the coordinators in the surrounding activity areas, and the campus personnel database, ensuring the accuracy of the collaborative team. Based on the accurate control of the warning event by the collaborative team, the safety level of the surrounding activity areas is fully considered, thus improving the accuracy of intelligent early warning of campus safety risks.

[0178] At this point, the system will retrieve the information of the person in charge of the activity area under warning from the previous steps. This person in charge is a key personnel responsible for the safety management or emergency response of the area. The system will verify the current status of the person in charge to ensure that they can respond and perform their duties. At the same time, the system will clarify the specific responsibilities of the person in charge, including command and coordination, resource allocation, and information communication.

[0179] The system collects information on collaborators in the surrounding activity areas from the campus personnel database. This information includes the person's name, position, contact information, and professional skills. Based on the type of warning event and the security level of the surrounding area, the system will match collaborators with the corresponding professional skills. For example, fire warnings require firefighters and medical rescue personnel, while personnel evacuation requires security personnel and volunteers.

[0180] Based on the leader's guidance and the professional skills of the collaborators, the system will build a collaborative team, which includes the leader, collaborators from different professions, and volunteers or assistants. After the team is built, the system will assign specific roles and tasks to each member, which are related to the member's professional skills, experience, and current status.

[0181] The system will establish effective communication channels for collaborative teams, such as walkie-talkies, text messages, and instant messaging software, to ensure real-time communication and collaboration among members. The system will also establish a collaborative process, including information transmission, decision-making, task execution, and feedback, to ensure that team members can work together efficiently.

[0182] Specifically, suppose a university's campus intelligent management system triggers a fire alarm in the library area and identifies the library as an activity area under alarm. At the same time, it assesses the safety level of the surrounding areas (study rooms, teaching buildings, and dormitories) and identifies the person in charge and collaborators.

[0183] The system retrieved the person in charge of the library fire early warning event—Director A of the campus security department—from the previous steps; the system verified Director A's status, confirming that he could respond immediately and fulfill his responsibilities; at the same time, the system clarified Director A's responsibilities, including directing and coordinating on-site emergency response work, allocating resources, and information communication.

[0184] The system collected information on coordinating personnel in the surrounding activity area from the campus personnel database, including firefighters, medical rescue personnel, and security personnel. Based on the type of fire warning and the safety level of the surrounding area, the system matched coordinating personnel with corresponding professional skills. Based on Director A's leadership and the professional skills of the coordinating personnel, the system constructed a collaborative team. This team included Director A as the leader, firefighters responsible for fire fighting and search and rescue, medical rescue personnel responsible for treating the injured, security personnel responsible for evacuation and maintaining order at the scene, and some volunteers or auxiliary personnel providing other necessary support.

[0185] The system establishes effective communication channels for the collaborative team, such as walkie-talkies and instant messaging software, to ensure real-time communication and collaboration among members. The system has developed a collaborative process, including Director A being responsible for overall command and decision-making, firefighters and medical rescue personnel carrying out firefighting and rescue tasks according to the situation on site, security personnel being responsible for guiding personnel evacuation and maintaining order on site, and volunteers or auxiliary personnel providing support such as material handling and information transmission. Team members transmit information in real time through communication channels and work together to respond to fire warning events.

[0186] Specifically, suppose the campus matching table shows entry three for library fire warning events:

[0187] Table 3, Campus Matching Table, contains entries related to library fire early warning events.

[0188] ,

[0189] When the library triggers a fire alarm, the system will automatically identify Director A as the person in charge based on the matching table, and search the campus personnel database for qualified firefighters, medical rescue personnel, and security personnel to form a collaborative team.

[0190] Please see Figure 8 , Figure 8 This is a schematic diagram of the structural composition of the intelligent early warning system for campus security risks in an embodiment of the present invention;

[0191] like Figure 8 As shown, an intelligent early warning system for campus security risks includes:

[0192] The job identifier module 21 is used to determine the corresponding job identifier based on the job information of each campus staff member and the corresponding personnel information.

[0193] Activity area module 22 is used to determine the activity areas of multiple campus personnel based on their job identifiers, the 3D model of the campus, and the time.

[0194] Security level module 23 is used to determine the security level of each activity area based on the location of each activity area, the type of activity in each activity area, and the density of people on campus.

[0195] The campus intelligent management database module 24 is used to determine the campus intelligent management database based on the security level of each activity area, the security matching coefficient of multiple adjacent activity areas, and deep learning of past campus security incidents.

[0196] The early warning module 25 is used to determine the activity area in the early warning state based on the analysis of the early warning signals of the campus in the campus intelligent management database;

[0197] The Collaborative Team Module 26 is used to determine the collaborative team based on the warning events of the activity area in the warning state, the security level of the surrounding activity areas, and the warning matching table.

[0198] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. An intelligent early warning method for campus safety risks, characterized in that, include: The corresponding job identifier is determined based on the job information of each campus staff member and the corresponding personnel information. The activity areas of multiple campus personnel are determined based on the identification of each position, the three-dimensional model of the campus, and the time. The security level of each activity area is determined based on its location, the type of activities within each area, and the density of people on campus. A campus intelligent management database is established based on the security levels of each activity area, the security matching coefficients of multiple adjacent activity areas, and deep learning of past campus security incidents. This includes: determining the activity areas to be managed based on the security levels of each activity area and preset security level thresholds; determining the security matching coefficients of multiple adjacent activity areas based on the matching of the activity areas to be managed and their corresponding adjacent activity areas; collecting a 3D model of the campus; determining past campus security incidents based on the 3D model, the activity areas to be managed, and their corresponding adjacent activity areas; performing deep learning on the security levels of each activity area, the security matching coefficients of multiple adjacent activity areas, and past campus security incidents; and establishing the campus intelligent management database based on the deep learning of the security levels of each activity area, the security matching coefficients of multiple adjacent activity areas, and past campus security incidents; and calculating the security matching coefficients between adjacent areas and the areas to be managed based on the security correlation between them. The campus intelligent management database identifies activity areas under warning status based on the analysis of campus warning signals. The collaborative team is determined based on the warning events in the activity area under warning status, the security level of the surrounding activity areas, and the warning matching table.

2. The intelligent early warning method for campus security risks according to claim 1, characterized in that, The process of determining the corresponding job identifier based on the job information of each campus staff member and the corresponding personnel information includes: The personnel information of each campus staff member is determined based on the campus roster and the personal database of campus personnel. The job information of each campus staff member is determined based on the campus staff database, the activity paths of each campus staff member, and the job responsibilities of each campus staff member. The corresponding job identifier is determined based on the job information of each campus staff member and the corresponding personnel information.

3. The intelligent early warning method for campus security risks according to claim 1, characterized in that, The process of determining the activity areas of multiple campus personnel based on job identifiers, a 3D model of the campus, and time includes: Multiple images of the campus were collected by drones taking dynamic pictures of the campus, and a three-dimensional model of the campus was determined based on the multiple images of the campus, the building distribution map of the campus, and the shape of the campus. Associate each job identifier with the 3D model of the campus, and record the current position of each job identifier in the 3D model of the campus; The first area parameters are determined based on the current location of various job identifiers and the current work content corresponding to each job identifier. The second area parameters are determined based on the current location and time of each job identifier. The activity areas of multiple campus personnel are determined based on the first area parameters, the second area parameters, and the three-dimensional model of the campus.

4. The intelligent early warning method for campus security risks according to claim 1, characterized in that, The determination of the security level of each activity area based on its location, the type of activity, and the density of people on campus includes: Mark the location of each activity area, and determine the multiple activities in each activity area by tracing the location of each activity area; In each activity area, the activity type is determined by deep learning based on multiple activity contents, corresponding campus personnel, and corresponding activity scope. The density of campus personnel is determined based on the area of ​​each activity zone, the distribution of corresponding campus personnel, and the number of campus personnel. The security level of each activity area is determined based on its location, the type of activities within each area, and the density of people on campus.

5. The intelligent early warning method for campus security risks according to claim 1, characterized in that, The process of determining the activity area under warning status in the campus intelligent management database based on the analysis of campus warning signals includes: Real-time monitoring of the campus intelligent management database; The campus intelligent management database outputs corresponding early warning signals, and the corresponding early warning information is determined by tracing the early warning signals.

6. The intelligent early warning method for campus security risks according to claim 5, characterized in that, The method of determining the activity area under warning status based on the analysis of campus warning signals in the campus intelligent management database also includes: Multiple activity areas are identified based on early warning information, and synchronous monitoring of multiple activity areas is triggered. Multiple activity warning features are output based on the synchronous monitoring of multiple activity areas. The activity areas under warning status are determined based on the location of the multiple activity warning features, the range of influence of the multiple activity warning features, and the spatial location of multiple campus personnel.

7. The intelligent early warning method for campus security risks according to claim 1, characterized in that, The process of determining the collaborative team based on the warning events in the activity area under warning status, the security level of surrounding activity areas, and the warning matching table includes: Multiple sub-activity events are output based on the monitoring of the activity area in the early warning state; The warning events for activity areas under warning status are determined based on multiple sub-events, the degree of injury to campus personnel, and the number of personnel in each campus. The surrounding activity areas are determined based on the location of the activity areas under warning status, and the corresponding safety level is determined by safety inspection of the surrounding activity areas.

8. The intelligent early warning method for campus security risks according to claim 7, characterized in that, The method of determining the collaborative team based on the warning events in the activity area under warning status, the security level of surrounding activity areas, and the warning matching table also includes: Collect the early warning matching table and associate the early warning events in the activity area under early warning status with the security levels of the surrounding activity areas into the early warning matching table; Based on the warning events in the activity area under warning status and the security level of the surrounding activity areas, determine the collaborating personnel in the surrounding activity areas; based on the warning events in the activity area under warning status and the warning matching table, determine the person in charge of the activity area under warning status. The collaborative team is determined based on the person in charge of the activity area under warning, the collaborators in the surrounding activity areas, and the campus personnel database.

9. An intelligent early warning system for campus safety risks, characterized in that, The intelligent early warning system for campus security risks is applied to the intelligent early warning method for campus security risks as described in any one of claims 1-8, wherein the intelligent early warning system for campus security risks includes: The job identifier module is used to determine the corresponding job identifier based on the job information of each campus staff member and the corresponding personnel information. The activity area module is used to determine the activity areas of multiple campus personnel based on their job identifiers, a 3D model of the campus, and the time. The security level module is used to determine the security level of each activity area based on its location, the type of activity in each activity area, and the density of people on campus. The campus intelligent management database module is used to determine the campus intelligent management database based on the security level of each activity area, the security matching coefficient of multiple adjacent activity areas, and deep learning of past campus security events. This includes: determining the activity areas to be managed based on the security level of each activity area and preset security level thresholds, and determining the security matching coefficient of multiple adjacent activity areas based on the matching of the activity areas to be managed and their corresponding adjacent activity areas; collecting a 3D model of the campus, and determining past campus security events based on the 3D model, the activity areas to be managed, and their corresponding adjacent activity areas; performing deep learning on the security level of each activity area, the security matching coefficient of multiple adjacent activity areas, and past campus security events, and determining the campus intelligent management database based on the deep learning of the security level of each activity area, the security matching coefficient of multiple adjacent activity areas, and past campus security events; and calculating the security matching coefficient between adjacent areas and the areas to be managed based on the security correlation between them. The early warning module is used to determine the activity areas under early warning status based on the analysis of early warning signals in the campus intelligent management database. The Collaborative Team module is used to determine collaborative teams based on the warning events in the activity area under warning status, the security level of surrounding activity areas, and the warning matching table.

Images