Intelligent campus intelligent security early warning system and method

Abstract

The invention discloses a smart campus intelligent security early warning system and method, and relates to the technical field of campus security, and the early warning system comprises a detection starting module, a mark matching module, a monitoring analysis module and an early warning adjustment module which are operated in sequence. The technical key points are as follows: the early warning system aims to monitor the pressure change near an enclosing wall in real time through a pressure plate, and performs personnel monitoring and identification in combination with a camera, so that intelligent campus security early warning is realized; the system can automatically judge the identity and behavior of a person, triggers early warning of different levels according to conditions, and intelligently dispatches security resources to ensure campus security. According to the scheme, the problems of low response speed, unreasonable personnel scheduling and the like in traditional campus security are effectively solved, and the emergency response efficiency and accuracy are improved; meanwhile, through application of the pre-estimation calculation model and the fuzzy calculation model, intelligence and refinement of security personnel dispatching are realized, the security requirement is ensured, and resource waste is avoided.

Description

Technical Field

[0001] The present invention relates to the field of campus security technology, and in particular to a smart campus intelligent security early warning system and method. Background Art

[0002] Campus security refers to a series of preventive measures and management systems taken to ensure the safety of teachers and students and maintain campus order. It covers many aspects, including but not limited to: access control management: by installing an intelligent access control system, control the people entering and leaving the campus, effectively prevent outsiders from entering at will, and ensure campus safety; video surveillance: install high-definition cameras in key areas of the campus to achieve all-weather, no-dead-angle monitoring, and promptly detect and handle abnormal situations; alarm system: set up emergency alarm devices, such as one-button alarm phones, alarm columns, etc., to ensure that they can respond quickly in an emergency and notify relevant departments; patrol inspections: arrange security personnel to patrol regularly, check the safety conditions of various parts of the campus, and prevent potential safety hazards.

[0003] However, for security processing actions near the campus walls, traditional security measures only use cameras for 24-hour continuous monitoring, and issue an early warning when someone is detected approaching or climbing over the campus wall. The early warning method is usually to send a signal to the remote monitoring end, and the response action of the monitoring end is to dispatch on-duty personnel to stop it. There are many problems with this early warning method. For each camera, in order to ensure the clarity of the monitoring process, the camera selected is a high-definition camera probe, and it is turned on 24 hours a day, which will cause excessive energy consumption and thus lead to excessively high detection costs. In addition, for the behavior of climbing over the campus wall, if there are too many people who engage in this behavior, it is necessary to manually control the number of security personnel to be dispatched. If the scheduling time is too long, it will lead to slow response speed in campus security. At the same time, there will be problems with unreasonable personnel scheduling. For example, if there are too many people climbing over the campus wall, only dispatching 2 security personnel will not be effective, resulting in poor overall security results. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the deficiencies in the prior art, the present invention provides a smart campus intelligent security warning system and method. The system monitors the pressure changes near the wall in real time through a pressure plate, and combines with a camera to monitor and identify personnel, thereby realizing intelligent campus security warning. The system can automatically determine the identity and behavior of personnel, and trigger different levels of warnings according to the situation. At the same time, it intelligently dispatches security resources to ensure campus safety, thereby solving the problems raised in the background technology.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] A smart campus intelligent security early warning system, the system comprising:

[0009] The detection start module monitors the pressure value output by the pre-laid pressure plate inside the target campus wall. When the instantaneous pressure value detected at any point exceeds the preset calibration threshold, the probe start signal is triggered;

[0010] When the probe at the corresponding position detects the target person, the continuous monitoring action is performed;

[0011] Otherwise, the probe at the corresponding position is turned off;

[0012] The tag matching module uses a deep learning algorithm to perform face recognition when the target person is detected by the probe at the corresponding position, and combines it with the campus database to determine whether the target person has entered the tag information;

[0013] If it is determined that the target person has no marked information, a first-level warning command will be issued;

[0014] Otherwise, the warning judgment mechanism is triggered;

[0015] The monitoring and analysis module executes the warning judgment mechanism and issues an audible and visual warning when it determines that the target person has marked information, until the probe at the corresponding position fails to detect the target person. When the audible and visual warning action continues, the module executes the behavior recognition strategy and combines computer vision technology to perform pre-recorded behavior analysis. When the target person is identified to have pre-recorded behavior, a secondary warning signal is issued and the marking information of the target person is output synchronously.

[0016] The early warning adjustment module executes the first-level adjustment strategy upon receiving the first-level early warning instruction;

[0017] Upon receiving the second-level warning instruction, execute the second-level adjustment strategy:

[0018] Input and extract the target personnel's tag information, build an estimated calculation model, generate an estimated dispatch value for security personnel, and dispatch personnel in the duty room according to the estimated dispatch value; based on the warning valuation calculated in the estimated calculation model, combined with the probe coordinates at the corresponding position and the distance value between the duty room, build a fuzzy calculation model, generate the response time, and complete the dispatch action within the response time when dispatching personnel in the duty room according to the estimated dispatch value.

[0019] Furthermore, the pre-laying position of the pressure plate is located on the inner side of the campus wall, and the distance between the pressure plate and the campus wall is S, and S>0.

[0020] Furthermore, the process of applying deep learning algorithms for face recognition is as follows:

[0021] Preprocessing: preprocess the facial image of the monitored target person;

[0022] Among them, preprocessing includes cropping, scaling, grayscale and denoising;

[0023] Feature extraction: Use convolutional neural network (CNN) to extract features from preprocessed face images;

[0024] Face recognition: compare the extracted feature vector with the face features already entered in the campus database, and determine whether they are the same person by calculating the similarity;

[0025] If the similarity exceeds the threshold, it is determined that the target person has the entered tag information;

[0026] If the similarity does not exceed the limit value, it is determined that the target person has no entered marking information.

[0027] Furthermore, the pre-recorded behavior at least includes wall climbing and climbing; the process of analyzing the pre-recorded behavior is as follows:

[0028] Use computer vision technology to perform behavioral analysis on the real-time video stream captured by the corresponding position probe;

[0029] By comparing the target person's behavior in the real-time video stream with the pre-recorded behavior, it is possible to identify whether the target person has pre-recorded behavior.

[0030] Furthermore, the first-level adjustment strategy is: obtain the coordinate information of the probe at the corresponding position, immediately dispatch all personnel in the duty room, and simultaneously notify the inspection personnel.

[0031] Furthermore, the tag information includes at least name, student number and class, and before building the estimation calculation model, the maximum number of people and grade in the target personnel are extracted.

[0032] Furthermore, the process of building the estimated calculation model is as follows:

[0033] According to the number of target personnel and the maximum grade, the early warning valuation calculation formula is established:

[0034]

[0035] In the formula, WV represents the warning valuation, N represents the number of target personnel, and M represents the maximum grade value of target personnel; the grade value ranges from 0 to 12;

[0036] Then, based on the early warning valuation and the known number of basic dispatches, the estimated dispatch value calculation formula is established:

[0037]

[0038] In the formula, EDV represents the estimated dispatch value, k represents the adjustment coefficient, which ranges from 0 to 1, and B represents the basic number of dispatched personnel.

[0039] Furthermore, the process of building a fuzzy computing model is as follows:

[0040] According to the warning estimate and the distance between the probe coordinates at the corresponding position and the duty room, the response time calculation formula is constructed:

[0041]

[0042] Modify formula (1) and add an upper limit constraint:

[0043]

[0044] In formulas (1) and (2), RT0 represents the response time set at the initial stage, and RT represents the response time finally generated.

[0045] A smart campus intelligent security early warning method comprises the following steps:

[0046] S1, monitoring the pressure value output by the pre-laid pressure plate inside the target campus wall, when the instantaneous pressure value detected at any point exceeds the preset calibration threshold, the probe start signal is triggered;

[0047] When the probe at the corresponding position detects the target person, the continuous monitoring action is performed;

[0048] Otherwise, the probe at the corresponding position is turned off;

[0049] S2. When the probe at the corresponding position detects the target person, the deep learning algorithm is used for face recognition, combined with the campus database to determine whether the target person has entered the tag information;

[0050] If it is determined that the target person has no marked information, a first-level warning command will be issued;

[0051] Otherwise, the warning judgment mechanism is triggered;

[0052] S3. Execute the warning judgment mechanism, and issue an audible and visual warning when it is determined that the target person has marked information, until the probe at the corresponding position does not detect the target person, and when the audible and visual warning action continues, execute the behavior recognition strategy, and perform pre-recorded behavior analysis in combination with computer vision technology. When the target person is identified to have pre-recorded behavior, a secondary warning signal is issued, and the marking information of the target person is output synchronously;

[0053] S4. Upon receiving the first-level warning instruction, execute the first-level adjustment strategy;

[0054] Upon receiving the second-level warning instruction, execute the second-level adjustment strategy:

[0055] Input and extract the target personnel's tag information, build an estimated calculation model, generate an estimated dispatch value for security personnel, and dispatch personnel in the duty room according to the estimated dispatch value; based on the warning valuation calculated in the estimated calculation model, combined with the probe coordinates at the corresponding position and the distance value between the duty room, build a fuzzy calculation model, generate the response time, and complete the dispatch action within the response time when dispatching personnel in the duty room according to the estimated dispatch value.

[0056] (III) Beneficial effects

[0057] The present invention provides a smart campus intelligent security early warning system and method, which has the following beneficial effects:

[0058] This solution uses pre-laid pressure plates to detect pressure values ​​in real time. When the instantaneous pressure value exceeds the preset calibration threshold, the probe at the corresponding position is accurately triggered to start, effectively avoiding false alarms and energy waste. At the same time, combined with the conditional triggering of high-definition camera probes, it not only ensures the clarity of the monitoring image, but also significantly reduces energy consumption, solving the high cost problem caused by the traditional 24-hour all-day probe opening.

[0059] This solution combines deep learning algorithms for face recognition, accurately distinguishes target personnel inside and outside the school, and conducts real-time monitoring and analysis of abnormal behaviors of personnel inside the school. Once an intrusion by an outsider or abnormal behavior of a person inside the school is detected, the system can quickly issue a warning signal of the corresponding level and simultaneously output the target person's tag information so that timely response measures can be taken. This improves the intelligence and refinement of campus security management and provides strong support for the construction of a smart campus.

[0060] This solution automatically executes different adjustment strategies according to the warning level to ensure the rational allocation and efficient use of security resources;

[0061] In case of a Level 1 warning, all on-duty personnel are dispatched immediately and patrol personnel are notified. The geographic information system is used to quickly locate and optimally dispatch security personnel to the scene of the incident. In case of a Level 2 warning, the estimated calculation model is used to comprehensively consider the number and grade of the target personnel, generate a reasonable estimated dispatch value, and complete the dispatch within the response time determined by the fuzzy calculation model.

[0062] This solution effectively solves the problems of slow response speed and unreasonable personnel scheduling in traditional campus security, and improves the efficiency and accuracy of emergency response. At the same time, through the application of predictive calculation models and fuzzy calculation models, it realizes the intelligent and refined dispatch of security personnel, which not only ensures safety needs but also avoids waste of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a modular schematic diagram of the early warning system in the present invention;

[0064] Figure 2 It is a schematic diagram of the configuration of the campus wall and the pressure plate in the present invention;

[0065] Figure 3 It is a schematic diagram of the overall process of the early warning method in the present invention. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] Embodiment 1:

[0068] See also Figure 1 to Figure 2 , this embodiment provides a smart campus intelligent security warning system; the system overview of this system is: it aims to monitor the pressure changes near the wall in real time through the pressure plate, and combine the camera to monitor and identify people, so as to realize intelligent campus security warning; the system can automatically determine the identity and behavior of people, and trigger different levels of warnings according to the situation, and intelligently dispatch security resources to ensure campus safety;

[0069] The early warning system includes various functional modules that run in sequence, namely: detection and start module, tag matching module, monitoring and analysis module, and early warning adjustment module; the entire early warning system conducts intelligent monitoring and early warning for campus wall climbing incidents, ensuring the safety of the campus;

[0070] Functional modules and descriptions

[0071] The detection start module monitors the pressure value output by the pre-laid pressure plate inside the target campus wall. When the instantaneous pressure value detected at any point exceeds the preset calibration threshold, the probe start signal is triggered and the probe at the corresponding position starts working;

[0072] When the probe at the corresponding position detects the target person, the continuous monitoring action is performed;

[0073] When the probe at the corresponding position fails to detect the target person, it indicates a false alarm, and the probe at the corresponding position is turned off;

[0074] Among them, refer to Figure 2 You can then see the location of the target campus's walls and its pressure plates;

[0075] The pre-laying position of the pressure plate is usually located inside the campus wall, and the distance from the campus wall is S. The value of S can be determined according to actual needs. For example, there is a drainage ditch close to the edge of the wall in the XX campus, so the pressure plate can be laid at intervals of the drainage ditch. At this time, the value of S is close to the maximum width of the drainage ditch.

[0076] For pressure plates:

[0077] They are usually part of security monitoring technology and are used to detect and alert intruders or escapees near fences or borders. Such devices are usually called "pressure sensors" or "pressure pads" and consist of a pad and several sensors installed under the pad, and can be integrated with security systems (such as alarm systems, monitoring systems or access control systems). The working principle of pressure sensors is based on pressure changes to detect whether an object (such as a person, animal or vehicle) is exerting pressure on it. When a person or object steps on the pressure plate, it triggers a signal, which can be transmitted to the central control unit, which in turn triggers an alarm or takes other security measures.

[0078] Burying pressure plates in the campus walls can effectively enhance campus security, especially in areas that require high security monitoring, such as school boundaries, around important facilities, or unattended areas at night; the pressure plates are part of the security system, and they usually need to be used in conjunction with other security equipment (such as cameras, lighting equipment, access control systems, etc.) to form a comprehensive and effective security monitoring system; at the same time, the installation and maintenance of such equipment also requires professional operation to ensure its normal operation and reliability;

[0079] The length of the pressure plate is limited, usually based on the shooting clarity limit of the corresponding probe. For example, a Class A camera can clearly shoot the facial image of a person within 300m, so the length of the pressure plate is 300m. The Class A camera is flush with one end of the pressure plate, and the monitoring range covers the corresponding pressure plate and the campus wall; each Class A camera corresponds to a pressure plate, and the position of the Class A camera is flush with the interval between adjacent pressure plates;

[0080] It should be noted that:

[0081] The preset calibration threshold is set manually, usually according to the minimum weight of students on campus. Some cats, dogs or other objects will not trigger the probe start signal, thereby completing targeted monitoring of students.

[0082] Each probe is not turned on 24 hours a day, but is turned on intermittently or under special circumstances according to the situation. In order to ensure the clarity of the monitoring process, high-definition camera probes are selected. Continuous opening will consume too much energy, which will lead to excessively high detection costs. Therefore, in order to meet the detection effect and control the detection cost, conditional triggering is used to ensure the effectiveness of security work while solving the cost problem.

[0083] Specifically, the above scheme is adopted to realize efficient and energy-saving security monitoring of the inner side of the campus wall;

[0084] This technical solution detects the pressure value in real time through the pre-laid pressure plate. When the instantaneous pressure value exceeds the preset calibration threshold, the probe at the corresponding position is accurately triggered to start, effectively avoiding false alarms and energy waste. At the same time, combined with the conditional triggering of the high-definition camera probe, it not only ensures the clarity of the monitoring image, but also significantly reduces energy consumption, solving the high cost problem caused by the traditional 24-hour all-day probe opening. In addition, the laying position and length of the pressure plate can be flexibly adjusted according to actual needs, and the coordination with the camera has formed a comprehensive and effective security monitoring system, enhancing the safety of the campus.

[0085] This technical solution not only improves the pertinence and effectiveness of security monitoring, but also fully considers economy and operability, effectively solves the technical problems in campus security monitoring, provides strong support for campus security prevention work, and ensures the harmony and stability of the campus environment and the safety of students.

[0086] Tag matching module:

[0087] When the probe at the corresponding position detects the target person, the deep learning algorithm is used for face recognition, combined with the campus database to determine whether the target person has entered the tag information;

[0088] If it is determined that the target person has no marking information, it means that the target person is an off-campus person, and a first-level warning instruction is issued;

[0089] If it is determined that the target person has marking information, it means that the target person is a person on campus, triggering the warning judgment mechanism;

[0090] Among them, the process of applying deep learning algorithm for face recognition is as follows:

[0091] Preprocessing: Preprocess the facial images of the monitored target persons, including cropping, scaling, graying and denoising, to improve recognition accuracy; Feature extraction: Use the convolutional neural network CNN to extract features from the preprocessed facial images and generate a unique feature vector representing the face; Face recognition: Compare the extracted feature vector with the facial features entered in the campus database and determine whether they are the same person by calculating the similarity;

[0092] If the similarity exceeds the threshold, they are considered to be the same person, indicating that the target person has entered tag information;

[0093] If the similarity does not exceed the threshold, it is determined that they are not the same person, which means that the target person has no entered tag information;

[0094] It should be noted that for non-school personnel, Figure 2 As shown in the figure, people outside the school will enter the campus from outside the campus wall, and people inside the school will go out from the campus by climbing over the campus wall. The early warning system completes the security early warning operation of the campus for these two situations and realizes smart campus management;

[0095] For the target person, it is at least one person, and usually several people;

[0096] Campus database comparison and judgment

[0097] Database construction: The campus database should contain the facial feature data of all campus personnel with entered tag information, as well as related tag information, where the tag information at least includes name, student number and class; Comparison judgment: The captured facial features are compared with the features in the database one by one to find the record with the highest similarity; Judgment result: No tag information: If the similarity does not exceed the defined value, that is, no matching record is found, the target person is judged to be an off-campus person; With tag information: If the similarity exceeds the defined value, that is, a matching record is found, the target person is judged to be an on-campus person;

[0098] In addition, it also includes ongoing operations, such as data updates: regularly updating the campus database to ensure that the facial feature data therein is consistent with the actual personnel; algorithm optimization: based on the actual application results, continuously optimizing the deep learning algorithm and face recognition model to improve recognition accuracy and efficiency. These are not explained in detail here.

[0099] Monitoring and analysis module:

[0100] Execute the warning judgment mechanism, and issue an audible and visual warning when it is determined that the target person has marked information, until the probe at the corresponding position does not detect the target person, and execute the behavior recognition strategy when the audible and visual warning action continues;

[0101] Combined with computer vision technology to analyze pre-recorded behaviors, when the target person is identified to have pre-recorded behaviors, a secondary warning signal is issued and the target person's tag information is output synchronously;

[0102] Among them, the sound and light warning method is carried out through the sound and light buzzer equipped on the corresponding probe, which plays a primary warning role; the purpose is to prevent school personnel from mistakenly entering the area near the campus wall and play a warning role;

[0103] In actual applications, if the students have no intention of climbing over the wall, they will leave immediately after hearing the sound and light warning. At this time, the probe at the corresponding position has not detected the target person, so the sound and light warning is turned off. However, some bold students may climb over the wall during campus sports events or when the noise of external interference is too loud. At this time, the effect of the sound and light warning is limited, so subsequent related strategies or actions are needed to play an effective security role.

[0104] The pre-recorded behaviors include at least wall climbing and scrambling;

[0105] The process for conducting a pre-recorded behavior analysis is as follows:

[0106] Use computer vision technology to perform behavioral analysis on the real-time video stream captured by the corresponding position probe;

[0107] Pre-recorded behavior analysis: The system pre-defines several behaviors that may pose security risks, such as climbing over walls and climbing. By comparing the target person's behavior in the real-time video stream with the pre-recorded behavior patterns, the system can identify whether the target person has these pre-recorded behaviors;

[0108] When the behavior recognition strategy confirms that the target person has pre-recorded behavior (such as climbing over a wall or climbing), a level 2 warning signal is immediately issued. Compared with the level 1 warning signal, there is no distinction between the two levels, but the distinction between the subsequent response and disposal actions. The warning signals of the same level correspond to different execution actions.

[0109] Secondary warning signals and marking information can be sent simultaneously in a variety of ways;

[0110] It should include at least one or a combination of SMS, APP push, email notification and campus broadcasting system to ensure that relevant information can be quickly conveyed to relevant personnel and security departments.

[0111] Specifically, the above solution is used to achieve efficient intelligent monitoring and early warning of the campus wall area;

[0112] The solution combines deep learning algorithms for face recognition, accurately distinguishing target personnel inside and outside the school, and conducting real-time monitoring and analysis of abnormal behaviors of personnel inside the school (such as climbing over walls). Once an intrusion by an outsider or abnormal behavior by a person inside the school is detected, the system can quickly issue an early warning signal of the corresponding level and simultaneously output the target person's tag information so that timely countermeasures can be taken.

[0113] This technical solution effectively solves the problems of high false alarm rate and slow response in traditional campus security systems, improves the intelligence and refinement of campus safety management, and provides strong support for the construction of smart campuses; at the same time, the combination of sound and light warnings and multiple warning signal sending methods ensures the timely transmission and effective response of warning information, further enhancing the campus's safety prevention capabilities.

[0114] Early warning adjustment module:

[0115] Upon receiving the first-level warning instruction, the first-level adjustment strategy is executed:

[0116] Obtain the coordinate information of the corresponding position probe, immediately dispatch all personnel in the duty room, and simultaneously notify the inspection personnel;

[0117] Among them, the first-level warning command indicates that there are outsiders invading the school. Therefore, for the number of dispatched personnel, since the danger of outsiders is higher than that of insiders, the data is as large as possible. The response time reserved for personnel is not an interval, but immediate, to avoid the situation where the danger increases due to adjustment or sorting actions;

[0118] To accurately obtain the geographical coordinates of the probe that triggered the first-level warning, this is usually achieved through the probe's built-in positioning function (using the built-in GPS) or the Geographic Information System (GIS) associated with the probe; the acquired coordinate information should be quickly and accurately transmitted to the responding personnel so that they can quickly locate the specific location of the incident;

[0119] For responders who need to go to the corresponding coordinates:

[0120] Usually, it is necessary to use geographic information system (GIS) and path planning algorithm to achieve the optimal scheduling of security personnel. In the scheduling process, commonly used path planning algorithms include Dijkstra algorithm and A* algorithm. Use the path planning algorithm to search for the optimal path from the starting point to the end point in the road network on campus. Among them, various factors can also be considered, such as the shortest distance, the shortest time, avoiding congested sections, etc., to meet the needs of different scenarios, which will not be elaborated here. The generated path is displayed to the security personnel in the form of a map, and detailed route information (such as action distance, estimated time, etc.) is provided. According to the actual situation and the feedback of the security personnel, the path is optimized and adjusted to ensure that the security personnel can reach the designated location smoothly.

[0121] Upon receiving the second-level warning instruction, execute the second-level adjustment strategy:

[0122] Input the target personnel's tag information, extract the maximum number and grade of the target personnel, build an estimated calculation model, generate the estimated dispatch value of the security personnel, and dispatch the personnel in the duty room according to the estimated dispatch value; based on the warning valuation calculated in the estimated calculation model, combined with the distance value between the probe coordinates of the corresponding position and the duty room, build a fuzzy calculation model, generate the response time, and complete the dispatch action within the response time when dispatching the personnel in the duty room according to the estimated dispatch value;

[0123] The marking information includes at least name, student number and class;

[0124] Extract the maximum value of the number of people and grade of the target personnel in the marked information (usually the system is adapted to the schools corresponding to junior high school and high school, so the grade numbers are from grade 7 to grade 12); it should be noted that: grade 7 corresponds to the first grade of junior high school; grade 8 corresponds to the second grade of junior high school; grade 9 corresponds to the third grade of junior high school; grade 10 corresponds to the first grade of senior high school; grade 11 corresponds to the second grade of senior high school; grade 12 corresponds to the third grade of senior high school;

[0125] In addition, the target person's marking information can also be synchronously matched and sent to the head teacher of the corresponding grade. For example, when the target person is 1, the name of the target person is XX, the student number is 32, and the class is Class 1, Grade 7, the marking information is sent to the academic affairs system for matching, and the head teacher of Class 1, Grade 7 is XX. The target person's marking information and the behavior of crossing the campus wall are packaged and sent to XX via SMS or other means, so that the head teacher can deal with the student in time and conduct subsequent ideological education or contact parents.

[0126] The process of building an estimated calculation model is as follows:

[0127] According to the number of target personnel and the maximum grade, the early warning valuation calculation formula is established:

[0128]

[0129] In the formula, WV represents the warning valuation, N represents the number of target personnel, and M represents the maximum grade value of the target personnel. Usually, the grade value ranges from 1 to 12, from elementary school to high school.

[0130] Logical explanation: The early warning valuation can be calculated based on the maximum number of people and grade level;

[0131] Taking into account the different effects of the number of people and the maximum grade on security needs, we can take their product and add an adjustment term to reflect the additional impact of the maximum grade on security needs; higher grades may mean more complex situations and higher risks, so more security personnel are needed;

[0132] The N×M part takes into account the basic combined effects of the number of people and the maximum value of the grade level;

[0133] The M(M-1) / 2 part is an additional adjustment term based on the maximum grade. It is half of a quadratic term about M and increases with the increase of M, reflecting the additional security needs brought about by the increase in grade.

[0134] Then, based on the early warning valuation and the known number of basic dispatches, the estimated dispatch value calculation formula is established:

[0135]

[0136] In the formula, represents the rounding up of ., EDV represents the estimated dispatch value, k represents the adjustment coefficient, and its value range is 0 to 1. It can be adjusted according to the actual situation, such as 0.1, 0.2, etc., to reflect the conversion ratio from the early warning estimate to the actual number of dispatched people; B represents the basic number of dispatched people, which is usually 1;

[0137] Logic description: Part I multiplies the warning estimate by an adjustment factor k and then rounds it up to ensure that the number of dispatched personnel is an integer; this adjustment factor can be set based on historical data or experience to reflect the relationship between the warning estimate and the actual dispatch demand; B is the basic dispatch number, which ensures that even if the warning estimate is very low, a certain number of security personnel will be dispatched;

[0138] In summary, through the two calculation formulas in the estimation calculation model, we can generate an estimated security personnel dispatch value based on the number of target personnel and the maximum grade. The first formula calculates the early warning valuation, taking into account the impact of the number of people and the maximum grade on security needs. The second formula calculates the final estimated dispatch value based on the early warning valuation and the basic dispatch number, ensuring that the dispatch number is a positive integer and is at least equal to the basic dispatch number. Such a model not only takes into account the complexity of the actual situation, but also ensures the rationality and operability of the dispatch number.

[0139] The process of building a fuzzy computing model is as follows:

[0140] According to the warning estimate and the distance between the probe coordinates at the corresponding position and the duty room, the response time calculation formula is constructed:

[0141]

[0142] In order to ensure that the response time does not exceed 30s, formula (1) is modified and an upper limit constraint is added:

[0143]

[0144] In formulas (1) and (2), RT0 represents the response time set at the primary level, and RT represents the final generated response time;

[0145] It should be noted that the distance between the probe coordinates and the duty room represents the navigation distance from the duty room to the probe coordinates, which is the actual distance rather than the straight-line distance;

[0146] Logical explanation: A basic response time is set to 10 seconds, considering that under normal circumstances, the duty room personnel need a certain amount of time to react after receiving the warning; the warning valuation WV reflects the impact of the number of target personnel and the maximum grade on security needs; we divide WV by 10 to ensure that the contribution of this part is not too large, while reflecting the positive impact of the warning valuation on the response time; the distance value D between the probe coordinates and the duty room has a direct impact on the response time; we divide D by 100 to keep the impact of the distance on the response time within a reasonable range; the min function is used to ensure that the response time does not exceed 30 seconds, in order to ensure that the response time does not exceed the set upper limit and avoid subsequent problems caused by untimely response;

[0147] Example:

[0148] Assume that the warning estimate WV = 150, and the distance between the probe coordinates and the duty room D = 200 meters;

[0149] Then apply the upper limit constraint: RT = min(27, 30) = 27;

[0150] Therefore, in this case, the response time is 27 seconds; this formula takes into account both the impact of the warning valuation on the response time and the fine-tuning effect of the distance value on the response time, while ensuring that the response time is within a reasonable range.

[0151] Specifically, the above scheme is used to achieve intelligent adjustment of campus safety warning and response;

[0152] The solution automatically implements different adjustment strategies according to the warning level (level 1 or level 2) to ensure the rational allocation and efficient use of security resources. When a level 1 warning is issued, all on-duty personnel are immediately dispatched and patrol personnel are notified. The geographic information system is used to quickly locate and optimally dispatch security personnel to the scene of the incident. When a level 2 warning is issued, the estimated calculation model is used to comprehensively consider the number and grade of the target personnel, generate a reasonable estimated dispatch value, and complete the dispatch within the response time determined by the fuzzy calculation model, and notify the relevant class teacher at the same time.

[0153] This solution effectively solves the problems of slow response speed and unreasonable personnel scheduling in traditional campus security, and improves the efficiency and accuracy of emergency response; at the same time, through the application of predictive calculation models and fuzzy calculation models, it realizes the intelligent and refined dispatch of security personnel, which not only ensures safety needs but also avoids waste of resources; in addition, the early warning information is sent to the class teacher at the same time, forming a safety prevention mechanism for home-school cooperation, which further improves the level and effectiveness of campus safety management.

[0154] Embodiment 2:

[0155] See also Figure 3 Based on Example 1, this embodiment also provides a smart campus intelligent security early warning method, including the following specific steps:

[0156] S1, monitoring the pressure value output by the pre-laid pressure plate inside the target campus wall, when the instantaneous pressure value detected at any point exceeds the preset calibration threshold, the probe start signal is triggered;

[0157] When the probe at the corresponding position detects the target person, the continuous monitoring action is performed;

[0158] Otherwise, the probe at the corresponding position is turned off;

[0159] S2. When the probe at the corresponding position detects the target person, the deep learning algorithm is used for face recognition, combined with the campus database to determine whether the target person has entered the tag information;

[0160] If it is determined that the target person has no marked information, a first-level warning command will be issued;

[0161] Otherwise, the warning judgment mechanism is triggered;

[0162] S3. Execute the warning judgment mechanism, and issue an audible and visual warning when it is determined that the target person has marked information, until the probe at the corresponding position does not detect the target person, and when the audible and visual warning action continues, execute the behavior recognition strategy, and perform pre-recorded behavior analysis in combination with computer vision technology. When the target person is identified to have pre-recorded behavior, a secondary warning signal is issued, and the marking information of the target person is output synchronously;

[0163] S4. Upon receiving the first-level warning instruction, execute the first-level adjustment strategy;

[0164] Upon receiving the second-level warning instruction, execute the second-level adjustment strategy:

[0165] Input and extract the target personnel's tag information, build an estimated calculation model, generate an estimated dispatch value for security personnel, and dispatch personnel in the duty room according to the estimated dispatch value; based on the warning valuation calculated in the estimated calculation model, combined with the probe coordinates at the corresponding position and the distance value between the duty room, build a fuzzy calculation model, generate the response time, and complete the dispatch action within the response time when dispatching personnel in the duty room according to the estimated dispatch value.

[0166] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. Before building the corresponding model for calculation, the required data will be dimensionless to ensure the smooth progress of the calculation process; when implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0167] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0168] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A smart campus intelligent security early warning system, characterized in that: The system includes: The detection start module monitors the pressure value output by the pre-laid pressure plate inside the target campus wall. When the instantaneous pressure value detected at any point exceeds the preset calibration threshold, the probe start signal is triggered; When the probe at the corresponding position detects the target person, the continuous monitoring action is performed; Otherwise, the probe at the corresponding position is turned off; The tag matching module uses a deep learning algorithm to perform face recognition when the target person is detected by the probe at the corresponding position, and combines it with the campus database to determine whether the target person has entered the tag information; If it is determined that the target person has no marked information, a first-level warning command will be issued; Otherwise, the warning judgment mechanism is triggered; The monitoring and analysis module executes the warning judgment mechanism and issues an audible and visual warning when it determines that the target person has marked information, until the probe at the corresponding position fails to detect the target person. When the audible and visual warning action continues, the module executes the behavior recognition strategy and combines computer vision technology to perform pre-recorded behavior analysis. When the target person is identified to have pre-recorded behavior, a secondary warning signal is issued and the marking information of the target person is output synchronously. The early warning adjustment module executes the first-level adjustment strategy upon receiving the first-level early warning instruction; Upon receiving the second-level warning instruction, execute the second-level adjustment strategy: Input and extract the target personnel's tag information, build an estimated calculation model, generate an estimated dispatch value for security personnel, and dispatch personnel in the duty room according to the estimated dispatch value; based on the warning valuation calculated in the estimated calculation model, combined with the probe coordinates at the corresponding position and the distance value between the duty room, build a fuzzy calculation model, generate the response time, and complete the dispatch action within the response time when dispatching personnel in the duty room according to the estimated dispatch value.

2. According to claim 1, the smart campus intelligent security early warning system is characterized by: The pre-laying position of the pressure plate is located on the inner side of the campus wall, and the distance between the pressure plate and the campus wall is S, and S>0.

3. According to claim 1, the smart campus intelligent security early warning system is characterized by: The process of applying deep learning algorithms for face recognition is as follows: Preprocessing: preprocess the facial image of the monitored target person; Among them, preprocessing includes cropping, scaling, grayscale and denoising; Feature extraction: Use convolutional neural network (CNN) to extract features from preprocessed face images; Face recognition: compare the extracted feature vector with the face features already entered in the campus database, and determine whether they are the same person by calculating the similarity; If the similarity exceeds the threshold, it is determined that the target person has the entered tag information; If the similarity does not exceed the limit value, it is determined that the target person has no entered marking information.

4. According to claim 1, the smart campus intelligent security early warning system is characterized by: The pre-recorded behaviors at least include wall climbing and climbing; the process of analyzing the pre-recorded behaviors is as follows: Use computer vision technology to perform behavioral analysis on the real-time video stream captured by the corresponding position probe; By comparing the target person's behavior in the real-time video stream with the pre-recorded behavior, it is possible to identify whether the target person has pre-recorded behavior.

5. According to claim 1, the smart campus intelligent security early warning system is characterized by: The first-level adjustment strategy is: obtain the coordinate information of the probe at the corresponding position, immediately dispatch all personnel in the duty room, and simultaneously notify the inspection personnel.

6. The smart campus intelligent security early warning system according to claim 1 is characterized by: The tag information includes at least name, student ID and class. Before building the estimation calculation model, the maximum number of people and grade among the target personnel are extracted.

7. The smart campus intelligent security early warning system according to claim 6 is characterized by: The process of building an estimated calculation model is as follows: According to the number of target personnel and the maximum grade, the early warning valuation calculation formula is established: In the formula, WV represents the warning valuation, N represents the number of target personnel, and M represents the maximum grade value of target personnel; the grade value ranges from 0 to 12; Then, based on the early warning valuation and the known number of basic dispatches, the estimated dispatch value calculation formula is established: In the formula, EDV represents the estimated dispatch value, k represents the adjustment coefficient, which ranges from 0 to 1, and B represents the basic number of dispatched personnel.

8. The smart campus intelligent security early warning system according to claim 7 is characterized by: The process of building a fuzzy computing model is as follows: According to the warning estimate and the distance between the probe coordinates at the corresponding position and the duty room, the response time calculation formula is constructed: Modify formula (1) and add an upper limit constraint: In formulas (1) and (2), RT0 represents the response time set at the initial stage, and RT represents the response time finally generated.

9. A smart campus intelligent security early warning method, using any system described in claims 1 to 8, characterized in that: The steps include: S1, monitoring the pressure value output by the pre-laid pressure plate inside the target campus wall, when the instantaneous pressure value detected at any point exceeds the preset calibration threshold, the probe start signal is triggered; When the probe at the corresponding position detects the target person, the continuous monitoring action is performed; Otherwise, the probe at the corresponding position is turned off; S2. When the probe at the corresponding position detects the target person, the deep learning algorithm is used for face recognition, combined with the campus database to determine whether the target person has entered the tag information; If it is determined that the target person has no marked information, a first-level warning command will be issued; Otherwise, the warning judgment mechanism is triggered; S3. Execute the warning judgment mechanism, and issue an audible and visual warning when it is determined that the target person has marked information, until the probe at the corresponding position does not detect the target person, and when the audible and visual warning action continues, execute the behavior recognition strategy, and perform pre-recorded behavior analysis in combination with computer vision technology. When the target person is identified to have pre-recorded behavior, a secondary warning signal is issued, and the marking information of the target person is output synchronously; S4. Upon receiving the first-level warning instruction, execute the first-level adjustment strategy; Upon receiving the second-level warning instruction, execute the second-level adjustment strategy: Input and extract the target personnel's tag information, build an estimated calculation model, generate the estimated dispatch value of security personnel, and dispatch personnel in the duty room according to the estimated dispatch value; Based on the warning valuation calculated in the estimation calculation model, combined with the distance value between the probe coordinates of the corresponding position and the duty room, a fuzzy calculation model is built to generate the response time. When the personnel in the duty room are dispatched according to the estimated dispatch value, the dispatch action is completed within the response time.

Images