Airport boundary grading early warning linkage method and system based on fuzzy logic reasoning

Abstract

The invention belongs to the technical field of airport boundary security and protection monitoring, and discloses an airport boundary grading early warning linkage method and system based on fuzzy logic reasoning. According to the method, an airport boundary is subjected to panoramic scanning to generate an infrared and visible light panoramic image, intrusion target detection and identification are performed on a boundary defense area in the infrared and visible light panoramic image through a multi-mode target detection algorithm based on deep learning, intrusion target early warning grade division is performed through a fuzzy logic algorithm, and an intrusion target early warning result is obtained. And carrying out early warning level priority judgment on the intrusion target, carrying out track labeling on a plurality of coordinate points closest to the intrusion target in the surrounding panoramic base map by using an OPENCV algorithm library, displaying the motion track of the current intrusion target, and judging the intention of the intrusion target. According to the invention, different airport environments are adapted through the rule base self-optimization capability, the manual patrol load is reduced while the high reliability is ensured, and the intelligent level and the safety prevention and control efficiency of the enclosure security and protection system are enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of airport perimeter security monitoring, and in particular relates to an airport perimeter graded early warning linkage method and system based on fuzzy logic reasoning. Background Art

[0002] In recent years, with the rapid development of the aviation industry, airport perimeter security systems are facing increasingly complex security challenges.

[0003] Traditional airport perimeter systems often rely on single technologies such as infrared beams, electronic fencing, and vibrating fiber optics. However, these technologies have significant drawbacks and are susceptible to interference from adverse weather conditions such as snow, rain, and heavy rain, resulting in a consistently high false alarm rate. This makes them unable to meet the ever-increasing demands of airport perimeter security systems. For example, perimeter systems based on traditional fiber optic sensors can experience hundreds of false alarms daily during typhoons.

[0004] Furthermore, relevant industry standards clearly stipulate that airport perimeter alarm systems must have a graded warning function, capable of accurately identifying different types of intrusion behaviors and implementing appropriate preventive measures based on the degree of danger. However, traditional airport perimeter systems often lack this alarm classification capability. When multiple targets are present within the perimeter, they are unable to effectively and intelligently classify intrusion targets to assist perimeter personnel in accurately locating the highest-level targets. Even some systems that do have alarm classification capabilities rely on fixed thresholds and pre-set rules, resulting in a relatively rigid static classification algorithm. For example, the system often defines a defense zone within 5 meters of the perimeter. This makes it impossible to effectively identify pre-intrusion behaviors such as throwing or loitering outside the perimeter, making it difficult to meet the actual needs of dynamic airport security. Therefore, a more intelligent and effective method and linkage system for intelligent graded warnings within airport perimeter systems is needed. Summary of the Invention

[0005] To overcome the problems existing in related technologies, the disclosed embodiments of the present invention provide a method and system for linking airport perimeter warnings based on fuzzy logic reasoning. This invention aims to address the existing technical issues of high false alarm rates, lack of intrusion risk grading, and static and rigid alarm grading using a single detection technology.

[0006] The technical solution is as follows: a method for linking airport perimeter graded warnings based on fuzzy logic reasoning, comprising the following steps:

[0007] S1: Panoramic scanning of the airport perimeter to generate infrared and visible light panoramic images, which are then sent to the outside world via the MQTT network protocol.

[0008] S2 receives infrared and visible light panoramic image data, and detects and identifies intruder targets in the perimeter defense area in the infrared and visible light panoramic images using a multimodal target detection algorithm based on deep learning. The detected and identified intruder target warning information is pushed, and the intruder target warning information includes the intruder target coordinates and intrusion feature information.

[0009] S3 receives and updates the infrared and visible light panoramic images sent in real time as the perimeter panoramic base map. It uses fuzzy logic algorithms to classify the sent intrusion target warning information into different levels, and then prioritizes the intrusion target warning levels. The prioritized intrusion target warning levels are then pushed to the server via HTTP.

[0010] S4, based on the priority judgment of the pushed intrusion target warning level, uses the OPENCV algorithm library to mark the trajectory of multiple coordinate points closest to the intrusion target in the surrounding panoramic base map, displays the current movement trajectory of the intrusion target, and judges the intrusion target intention.

[0011] In step S1, the dual-band panoramic monitoring subsystem performs a panoramic scan of the airport perimeter to generate infrared and visible light panoramic images. The dual-band panoramic monitoring subsystem includes multiple panoramic scanning monitoring devices, each of which is installed at intervals and simultaneously mounted with the dome camera on the cross arm of a pole mounted inside the airport perimeter. The pole is installed. By continuously rotating 360 degrees horizontally, each device generates a 360-degree panoramic image at fixed time intervals. The specific steps are as follows:

[0012] Step 1: Establish a model for the weakening mechanism of interference light disturbing the grayscale image and determine the optimization framework;

[0013] Step 2: constructing data items for modeling light disturbance;

[0014] Step 3: Select image prior terms and combine them with data terms to construct a non-optimal feedback model for grayscale image disturbance caused by interference light.

[0015] Step 4: Use iterative reweighted second-order iterative optimization method to obtain a clear panoramic image.

[0016] In step 1, a model for the weakening mechanism of interfering light perturbing the grayscale image is established, and an optimization framework is determined, including:

[0017] (1) Establish a nonlinear weakening model for the grayscale image disturbed by interference light, and the expression is:

[0018]

[0019] Where, is the vector of the blurred image that has not been polluted by grayscale, D is the known blur kernel, u is the vector of the original clear image, f is the vector of the final weakened image, and W is the process of the image being polluted by the disturbing light;

[0020] (2) According to the weakening model, based on the maximum a posteriori probability framework, data fidelity and image prior are combined into a unified image restoration framework. The non-optimal feedback problem corresponds to an optimization problem, which is expressed as:

[0021]

[0022] Where, is the image vector in the unified image restoration framework, C(u|f) is the posterior probability;

[0023] According to the Bayesian formula, we get:

[0024] C(u|f)∝C(f|u)C(u)

[0025] Where C(f|u) is the likelihood probability and C(u) is the prior probability;

[0026] Combining the above two formulas, the equivalent transformation is:

[0027]

[0028] Where, is the image vector after equivalent transformation;

[0029] The following optimization framework is obtained:

[0030]

[0031] In the formula, ψ(Du-f) is the data item to be constructed, is the image prior item to be designed;

[0032] In step 2, the data items for modeling the light disturbance are constructed, including:

[0033] (1) Select the following nonlinear function to eliminate the influence of disturbance grayscale:

[0034]

[0035] Where h and l are function parameters, h controls the nonlinearity of the function at the cutoff point, l controls the cutoff point, D(u) is the blur kernel after removing the influence of the disturbance grayscale, and e is the Reynolds number.

[0036] (2) Combine the nonlinear function of L1 norm and construct the following data items:

[0037]

[0038] Where i is the image pixel index, f is the vector of the final weakened image, S(u) is the data item of the image vector, and Q is the nonlinear function;

[0039] In step 3, a non-optimal feedback model for grayscale image disturbance caused by interference light is constructed by combining the data items, including:

[0040] The expression of the prior term constructed based on the image gradient sparse prior is:

[0041]

[0042] Where, L w is the sparse prior value of the image gradient of the w-th image node, is the image gradient sparse prior non-optimal feedback, L v is the sparse prior value of the image gradient of the vth image node;

[0043] In step 4, the iterative reweighted second-order iterative optimization method includes:

[0044] The non-optimal feedback model is transformed into a weighted second-order iterative optimization method as shown below:

[0045]

[0046] Where λ u is the regularization term weight, A is a weighted second-order iterative optimization method. w is the image gradient sparse prior value weight diagonal matrix in the w-th image node deformation, D w is the blur kernel in the deformation of the w-th image node, A v is the diagonal matrix of image gradient sparse prior weights in the v-th image node deformation, D v is the blur kernel in the deformation of the vth image node.

[0047] In step S2, receiving infrared and visible light panoramic image data includes:

[0048] The AI ​​intelligent analysis center subscribes to MQTT events to receive infrared and visible light panoramic images sent by the dual-band panoramic monitoring subsystem. Using a multimodal target detection algorithm based on deep learning, it detects and identifies intruder targets within the perimeter defense areas in the infrared and visible light panoramic images, and pushes warning information about detected and identified intruder targets to the intelligent alarm classification processing center.

[0049] Intrusion target detection includes target type, target behavior, target distance, and target residence time. Target type includes characteristics of people, vehicles, animals, and drones; target behavior includes characteristics of stationary, wandering, climbing, and throwing; target distance is the vertical distance between the target and the virtual perimeter in the panoramic image; the closer the distance, the higher the risk of the intrusion target; and target residence time is the time the intrusion target remains within the perimeter defense zone.

[0050] The deep learning-based multimodal target detection algorithm includes: first generating event codes based on the UUID standard and attaching spatiotemporal coordinate tags; then extracting multispectral fusion feature vectors, including RGB spectrum distribution, infrared radiation intensity, and morphological profile parameters; synchronously activating the timing analysis engine, combining the device's historical image library to reconstruct motion trajectories and match behavior patterns, and establish a three-dimensional spatial motion situation model of the target; and finally outputting structured feature dataset parameters, including target type, target behavior, target distance, and target residence time.

[0051] Establishing a three-dimensional target motion situation model includes:

[0052]

[0053] Where X, T, K, and t are target type motion values, target behavior motion trend motion values, target distance, and target retention time, respectively; G is the target motion matching parameter; J is the multispectral fusion feature sparseness; z is the vertical target motion area; F is the multispectral fusion feature vector; x is the horizontal target motion area; y is the vertical target motion area; and H is the target motion area. z is the vertical target trending motion coefficient, and I is the target motion trajectory radius.

[0054] In step S3, the fuzzy logic algorithm is used to classify the intrusion target warning levels and to determine the priority of the intrusion target warning levels, including:

[0055] S301, fuzzy design of invasion target characteristic variables;

[0056] S302, design of intrusion alarm fuzzy rule base;

[0057] S303, fuzzy reasoning and defuzzification of intrusion alarm features;

[0058] S304, the intelligent alarm classification processing center saves the intrusion target warning level obtained by fuzzy algorithm reasoning into the database, and pushes the intrusion target coordinate point and the intrusion target warning level to the linkage control center through the HTTP protocol.

[0059] In step S301, the fuzzy design of the intrusion target characteristic variables includes:

[0060] The target type membership function expression is:

[0061]

[0062] Where μ type is the target type membership, ranging from [0,1]; ω type is the target type risk weight, ranging from [0,1]; reliability is the target behavior confidence;

[0063] The target behavior membership function expression is:

[0064]

[0065] Where μ behavior is the target behavior membership, ranging from [0,1]; ω behavior is the risk weight of the target behavior, ranging from [0,1];

[0066] The target distance membership function expression is:

[0067]

[0068] Where μ distance is the target distance membership, ranging from [0, 1]; d is the vertical distance between the target and the boundary wall, m is the center point, m = 5; σ is the standard deviation, σ = 5;

[0069] The target residence time membership function expression is:

[0070]

[0071] Where μ time is the target residence time membership, ranging from [0,1]; k is the slope coefficient, T mid is the midpoint time, T crit is the critical time;

[0072] In step S302, the intrusion alarm fuzzy rule base is designed, including:

[0073] The fuzzy rule base is designed to meet the following rules:

[0074] (1) The fuzzy rule base should cover all input combination states;

[0075] (2) Full spatial division of the four characteristic variables of the airport perimeter intrusion targets: target type, target behavior, target distance, and target residence time;

[0076] (3) Use natural language to describe the rules, that is, the IF-THEN structure, which is expressed as follows:

[0077] IF(μ type ∈ Ltype )AND(μ behavior ∈L behavior )AND(μ distance ∈L distance )AND(μ ditance

[0078] ∈L time ) THEN Warning Level = C m

[0079] Among them, L variable is the fuzzy level of characteristic variable, C m The ultra-high / high / medium / low warning level output by the rule; L type is the fuzzy level of the characteristic variable of the target type, L behavior is the fuzzy level of the characteristic variable of the target behavior, L distance is the fuzzy level of the characteristic variable of target distance, L time is the fuzzy level of the characteristic variable of target residence time;

[0080] (4) Fuzzy rule strength calculation should be performed in accordance with the AND calculation method to calculate the rule strength information that the intrusion target characteristic variables meet. The expression is:

[0081] μ rulei =min(μ type ,μ behavior ,μ distance ,μ time )×ω i

[0082] Where μ rulei Adapt the activation strength of the i-th rule to the intrusion target, ω i is the rule weight of the i-th rule.

[0083] In step S303, fuzzy reasoning and defuzzification of intrusion alarm features include:

[0084] Based on fuzzy feature variables and fuzzy rule base, and taking the intrusion target feature variables as input, the method performs reasoning according to the steps of feature variable fuzzification, rule strength evaluation, rule consequence aggregation and defuzzification, and finally obtains the intrusion target warning level.

[0085] The rule strength evaluation is to perform a logical operation on each rule antecedent according to the fuzzy rule strength calculation formula based on the predefined fuzzy rule base to obtain the activation strength of the rule consequent;

[0086] The rule consequent aggregation combines the fuzzy sets of all activated rule consequents into a comprehensive output fuzzy set, using the maximum method MAX according to arg max(μrulei ) method, the results of multiple rules are superimposed and merged, μ rule =max(μ rule1 ,μ rule2 ,μ rule3 …μ rulen ) derives the alarm rule strength value, and then dynamically divides it according to the activation strength threshold maintained by the system. Common methods of dynamic optimization include manual configuration, data-driven optimization, and real-time adaptation;

[0087] Manual configuration involves manually adjusting the threshold range through the management interface; data-driven optimization involves automatically fitting the optimal threshold through machine learning based on historical intrusion event data to adapt to the environmental differences of different airports;

[0088] Real-time adaptation includes: combining real-time weather and passenger flow data to dynamically scale the threshold range; the warning level activation intensity range includes four levels: ultra-high [0.8, 1], high [0.6, 0.8], medium [0.3, 0.6], and low [0.0, 0.3].

[0089] In step S4, the OPENCV algorithm library is used to mark the trajectories of the multiple coordinate points closest to the intrusion target in the surrounding panoramic base map, display the current movement trajectory of the intrusion target, and judge the intrusion target intention, including:

[0090] The linkage control center can automatically identify and lock the most threatening target based on the warning level of the intrusion target; it can track the intrusion target by mapping the intrusion target coordinate point in the panoramic base map of the perimeter to the horizontal rotation angle P, pitch angle T and zoom factor Z of the ball camera in real time; at the same time, the linkage control center obtains the ball camera information bound to the current panoramic scanning device from the database, and builds a three-dimensional space coordinate conversion model based on the mapping relationship between the coordinate system of the panoramic base map of the perimeter and the physical coordinate system of the ball camera, automatically controlling the ball camera to turn towards the intrusion target direction for video tracking confirmation.

[0091] Furthermore, the linkage control center calculates the PTZ parameters of the ball camera based on the coordinates of the intrusion target (PositionX, PositionY), the vertical field of view angle ForVer, and the pitch angle DangleVer of the periscopic device. The specific mapping process is as follows:

[0092] The horizontal angle P uses a polar coordinate mapping algorithm to convert the target pixel coordinates (PositionX / ImgWidth) into the horizontal rotation angle of the dome camera:

[0093]

[0094] Where ImgWidth and ImgHeight are the pixel width and height of the panorama, θoffset The initial angle calibration offset between the scanning device and the ball camera is pre-measured and stored by the device calibration program. When the scanning device and the ball camera are at the same angle θ offset =0;

[0095] The vertical pitch T is calculated by combining the vertical field of view angle ForVer and the pitch angle DangleVer of the dome camera using a triangular geometry algorithm:

[0096]

[0097] In the formula, the negative sign indicates the engineering convention that the downward direction of the ball camera is the positive direction;

[0098] The zoom factor Z is:

[0099]

[0100] Where Z min ,Z max They are respectively the minimum and maximum zoom ratios preset for the speed dome camera.

[0101] Another object of the present invention is to provide an airport perimeter graded warning linkage system based on fuzzy logic reasoning, which implements the airport perimeter graded warning linkage method based on fuzzy logic reasoning, and includes:

[0102] A dual-band panoramic monitoring subsystem is used to scan the airport perimeter to generate infrared and visible light panoramic images, which are then transmitted via the MQTT network protocol.

[0103] The AI ​​intelligent analysis center receives infrared and visible light panoramic image data and uses a multimodal target detection algorithm based on deep learning to detect and identify intruder targets in the perimeter defense areas of the infrared and visible light panoramic images. It also pushes warning information about detected and identified intruder targets, which includes the target coordinates and intrusion characteristics.

[0104] The intelligent alarm classification processing center is used to receive and update the infrared and visible light panoramic images sent in real time as the perimeter panoramic base map. It uses fuzzy logic algorithms to classify the intrusion target warning information sent, and makes priority judgments on the intrusion target warning levels. The priority judgment of the intrusion target warning levels is pushed through the HTTP protocol.

[0105] The linkage control center is used to judge the warning level of the intrusion target after pushing the priority, use the OPENCV algorithm library to mark the trajectory of the multiple coordinate points closest to the intrusion target in the panoramic base map of the perimeter, display the movement trajectory of the current intrusion target, and judge the intention of the intrusion target.

[0106] Combining all the above technical solutions, the beneficial effects of the present invention are as follows: the present invention combines the panoramic images generated by the 360-degree scanning equipment, uses AI to intelligently identify the characteristics and behaviors of intruders, and focuses on using fuzzy logic algorithms for reasoning operations to achieve intelligent graded early warning of perimeter intrusion targets and dome-camera linkage tracking and confirmation functions, effectively reducing the false alarm rate and improving the accuracy and response speed of airport perimeter security alarms.

[0107] This invention builds an intelligent multi-feature data fusion mechanism based on fuzzy logic reasoning, effectively reducing false alarms caused by environmental interference. The system employs an adaptive hierarchical early warning strategy, transcending the limitations of traditional fixed thresholds. It dynamically determines threat levels through multi-dimensional feature analysis and real-time reasoning, and integrates dome cameras, audio and video equipment to form a hierarchical response closed loop, significantly improving emergency response efficiency and security accuracy. The self-optimization capabilities of the rule base adapt to different airport environments, reducing the manual inspection burden while ensuring high reliability, and enhancing the intelligence level and safety control effectiveness of the perimeter security system. BRIEF DESCRIPTION OF THE DRAWINGS

[0108] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure;

[0109] Figure 1 This is a schematic diagram of an airport perimeter graded warning linkage system based on fuzzy logic reasoning provided by an embodiment of the present invention;

[0110] Figure 2 This is a schematic diagram of the airport perimeter graded warning linkage method based on fuzzy logic reasoning provided by an embodiment of the present invention;

[0111] Figure 3 This is a flow chart of the logical reasoning of target features by the fuzzy logic algorithm provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0112] To make the above-mentioned objects, features, and advantages of the present invention more readily apparent, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings. The following description sets forth numerous specific details to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways than those described herein, and those skilled in the art may make similar modifications without departing from the scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0113] The innovation of this invention lies in: it breaks through the "single technology detection and fixed threshold classification" mode of the traditional perimeter system, improves detection accuracy through multimodal perception (dual-band panoramic monitoring technology + deep learning multimodal target detection), realizes dynamic classification through fuzzy logic (multi-dimensional feature fuzzy modeling + full space coverage fuzzy rule library), and optimizes response efficiency through intelligent linkage (panoramic base map trajectory visualization + three-dimensional coordinate mapping and ball camera linkage), forming a complete intelligent security closed loop of "detection-classification-response", which significantly improves the safety control efficiency and intelligence level of airport perimeters.

[0114] In Example 1, the airport perimeter graded warning linkage method based on fuzzy logic reasoning provided by the embodiment of the present invention performs AI intelligent analysis on the panoramic image scanned around the perimeter to confirm the characteristic factors such as the type, behavior, distance, and detention time of the intruder target, and then calculates the characteristic membership through the fuzzy logic algorithm and performs early warning graded management, and links the ball camera for video tracking. Figure 1 shown.

[0115] Figure 2 This is the principle of the airport perimeter graded warning linkage method based on fuzzy logic reasoning provided by an embodiment of the present invention.

[0116] Specifically, the airport perimeter graded warning linkage method based on fuzzy logic reasoning provided by the embodiment of the present invention includes:

[0117] S1: Panoramic scanning of the airport perimeter to generate infrared and visible light panoramic images, which are then sent to the outside world via the MQTT network protocol.

[0118] For example, the dual-band panoramic surveillance subsystem includes multiple panoramic scanning devices, each spaced apart and mounted alongside a dome camera on the crossbar of a pole mounted inside the airport perimeter. By continuously rotating 360 degrees horizontally, each device generates a 360-degree panoramic image at regular intervals.

[0119] MQTT is a lightweight, publish-subscribe based application layer messaging protocol suitable for resource-constrained devices and low-bandwidth, high-latency or unstable network environments.

[0120] The circumferential scanning monitoring device generates a 360-degree panoramic image at fixed time intervals, including:

[0121] Step 1: Establish a model for the weakening mechanism of interference light disturbing the grayscale image and determine the optimization framework;

[0122] Step 2: constructing data items for modeling light disturbance;

[0123] Step 3: Select image prior terms and combine them with data terms to construct a non-optimal feedback model for grayscale image disturbance caused by interference light.

[0124] Step 4: Use iterative reweighted second-order iterative optimization method to obtain a clear panoramic image.

[0125] Exemplarily, in step 1, establishing a blurred image reduction mechanism model and determining an optimization framework specifically include:

[0126] (1) Establish a nonlinear weakening model for the grayscale image disturbed by interference light, and the expression is:

[0127]

[0128] Where, is the vector of the blurred image that has not been polluted by grayscale, D is the known blur kernel, u is the vector of the original clear image, f is the vector of the final weakened image, and W is the process of the image being polluted by the disturbing light;

[0129] (2) According to the weakening model, based on the maximum a posteriori probability framework, data fidelity and image prior are combined into a unified image restoration framework. The non-optimal feedback problem corresponds to an optimization problem, which is expressed as:

[0130]

[0131] Where, is the image vector in the unified image restoration framework, C(u|f) is the posterior probability;

[0132] According to the Bayesian formula, we get:

[0133] C(u|f)∝C(f|u)C(u)

[0134] Where C(f|u) is the likelihood probability and C(u) is the prior probability;

[0135] Combining the above two formulas, the equivalent transformation is:

[0136]

[0137] Where, is the image vector after equivalent transformation;

[0138] The following optimization framework is obtained:

[0139]

[0140] In the formula, ψ(Du-f) is the data item to be constructed, is the image prior term to be designed, i.e., the regularization term, which constrains the clear panoramic image to conform to the manually designed statistical prior distribution, corresponding to the -logC(u) term;

[0141] In step 2, the method of constructing data items to model light disturbance specifically includes:

[0142] (1) Select the following nonlinear function to eliminate the influence of disturbance grayscale:

[0143]

[0144] Where h and l are function parameters, h controls the nonlinearity of the function at the cutoff point, l controls the cutoff point, D(u) is the blur kernel after removing the influence of the disturbance grayscale, and e is the Reynolds number.

[0145] (2) Combine the nonlinear function of L1 norm and construct the following data items:

[0146]

[0147] Where i is the image pixel index, f is the vector of the final weakened image, S(u) is the data item of the image vector, and Q is the nonlinear function;

[0148] In step 3, a non-optimal feedback model for grayscale image disturbance caused by interference light is constructed by combining the data items, including:

[0149] The expression of the prior term constructed based on the image gradient sparse prior is:

[0150]

[0151] Where, L w is the sparse prior value of the image gradient of the w-th image node, is the image gradient sparse prior non-optimal feedback, L v is the sparse prior value of the image gradient of the vth image node;

[0152] In step 4, the iterative reweighted second-order iterative optimization method includes:

[0153] The non-optimal feedback model is transformed into a weighted second-order iterative optimization method as shown below:

[0154]

[0155] Where λ u is the regularization term weight, A is a weighted second-order iterative optimization method. w is the image gradient sparse prior value weight diagonal matrix in the w-th image node deformation, D w is the blur kernel in the deformation of the w-th image node, Av is the diagonal matrix of image gradient sparse prior weights in the v-th image node deformation, D v is the blur kernel in the deformation of the vth image node.

[0156] S2 receives infrared and visible light panoramic image data, and detects and identifies intruder targets in the perimeter defense area in the infrared and visible light panoramic images using a multimodal target detection algorithm based on deep learning. The detected and identified intruder target warning information is pushed, and the intruder target warning information includes the intruder target coordinates and intrusion feature information.

[0157] The AI ​​intelligent analysis center receives infrared and visible light panoramic images sent by the dual-band panoramic monitoring subsystem by subscribing to MQTT events;

[0158] Exemplarily, the multimodal target detection algorithm based on deep learning is an algorithm that can perform target detection and analysis on infrared and visible light panoramic images. Through detection and analysis, characteristic information such as the target type, target behavior, target distance, and target residence time of the intrusion target contained in the image is obtained, thereby providing basic target feature variables for subsequent fuzzy logic reasoning algorithms.

[0159] For example, the AI ​​intelligent analysis center can identify and set up a virtual airport perimeter wall in a panoramic image, establish a core defense zone 5 meters inside and outside the virtual perimeter wall, and set up a warning zone 20 meters inside and outside the perimeter wall. It then detects characteristic factors such as target type, target behavior, target distance, and target residence time of all intruders within the perimeter defense zone, and pushes the target coordinates and intrusion characteristics of the intruder target warning information to the intelligent alarm classification processing center.

[0160] Exemplarily, intrusion target detection includes target type, target behavior, target distance, and target residence time. Target type includes characteristics such as people, vehicles, animals, and drones; target behavior includes characteristics such as stationary, wandering, climbing, and throwing; target distance is the vertical distance between the target and the virtual perimeter in the panoramic image; a closer distance indicates a higher risk; and target residence time is the time the target remains within the perimeter, generally in seconds; a longer residence time indicates a higher risk.

[0161] An exemplary multimodal target detection algorithm based on deep learning includes: first generating event codes based on the UUID standard and attaching spatiotemporal coordinate tags; then extracting multispectral fusion feature vectors, including RGB spectrum distribution, infrared radiation intensity and morphological contour parameters; synchronously activating the timing analysis engine, combining the device historical image library to reconstruct the motion trajectory and match the behavior pattern, and establish a three-dimensional spatial motion situation model of the target; and finally outputting the structured feature data set parameters, including target type, target behavior, target distance and target residence time.

[0162] Establishing a three-dimensional target motion situation model includes:

[0163]

[0164] Where X, T, K, and t are target type motion values, target behavior motion trend motion values, target distance, and target retention time, respectively; G is the target motion matching parameter; J is the multispectral fusion feature sparseness; z is the vertical target motion area; F is the multispectral fusion feature vector; x is the horizontal target motion area; y is the vertical target motion area; and H is the target motion area. z is the vertical target trending motion coefficient, and I is the target motion trajectory radius.

[0165] S3 receives and updates the infrared and visible light panoramic images sent in real time as the perimeter panoramic base map. It uses fuzzy logic algorithms to classify the sent intrusion target warning information into different levels, and then prioritizes the intrusion target warning levels. The prioritized intrusion target warning levels are then pushed to the server via HTTP.

[0166] The intelligent alarm classification processing center subscribes to MQTT events, and receives and updates the infrared and visible light panoramic images sent by the dual-band panoramic monitoring subsystem in real time as the perimeter panoramic base map. When the AI ​​intelligent analysis center identifies an intrusion target, the intelligent alarm classification processing center receives the intrusion target warning information sent by the AI ​​intelligent analysis center, classifies the intrusion target warning level using the fuzzy logic algorithm, and makes a priority judgment on the intrusion target warning level. The priority judgment of the intrusion target warning level is then pushed to the linkage control center via the HTTP protocol.

[0167] For example, in the process of determining the priority of the intrusion target warning levels, the intrusion target warning level with the highest risk can be obtained.

[0168] Exemplarily, the fuzzy logic algorithm is an algorithm based on fuzzy set theory and fuzzy reasoning. It is a mathematical tool for processing uncertainty and fuzzy information. It converts precise input quantities into fuzzy quantities through membership functions, performs logical reasoning based on the fuzzy rule base, and finally outputs decision values ​​through defuzzification. Its core processes include: fuzzification (mapping precise input quantities to fuzzy sets), establishing a fuzzy rule base (defining the logical relationship between input and output variables), fuzzy reasoning (performing fuzzy logic operations such as AND / OR), and defuzzification (converting fuzzy outputs into precise values). The application process of the fuzzy logic algorithm in the airport perimeter system for warning level classification is as follows:

[0169] S301, fuzzy design of invasion target characteristic variables;

[0170] The target type membership function expression is:

[0171]

[0172] Where μ type is the target type membership, ranging from [0,1]; ω type is the target type risk weight, ranging from [0,1]; reliability is the target behavior confidence;

[0173] The target type risk level can be dynamically set based on actual data and expert opinions and stored in the system database. For example, the risk level for people is 1.0, and the risk level for cars is 0.8. Reliability is the confidence level of the target type, ranging from [0, 100]. This is determined by analyzing the target type characteristics through the AI ​​Intelligent Analysis Center.

[0174] The target behavior membership function expression is:

[0175]

[0176] Where μ behavior is the target behavior membership, ranging from (0,1]; ω behavior is the risk weight of the target behavior, ranging from [0,1];

[0177] The target behavior risk level can be dynamically set based on actual data and expert opinions and stored in the system database. For example, the climbing risk level is 1.0, the throwing risk level is 0.7, etc. Reliability is the confidence level of the target behavior, ranging from [0,100]. This is determined by analyzing the target behavior characteristics through the AI ​​Intelligent Analysis Center.

[0178] The target distance membership function expression is:

[0179]

[0180] Where, μ distance is the membership degree of the target distance, with a range of [0, 1]; d is the vertical distance from the target to the perimeter wall, m is the center point, m = 5; σ is the standard deviation, σ = 5;

[0181] d is the vertical distance from the target to the perimeter wall. When 0 ≤ d ≤ 5 meters, the membership degree of the target distance is 1; when d > 20 meters, the membership degree of the target distance is 0; when 5 < d ≤ 20 meters, the membership degree of the target distance decreases non-linearly with the increase of the distance. According to the actual situation of the airport perimeter, the Gaussian function is used as the membership function of the warning area.

[0182] The expression of the membership function of the target residence time is:

[0183]

[0184] Where, μ time is the membership degree of the target residence time, with a range of [0, 1]; k is the slope coefficient, k = 0.3; T mid is the midpoint time, T mid = 10; T crit is the critical time, T crit = 20.

[0185] According to the non-linear increase of the residence time of the intrusion target in the perimeter defense area, when it exceeds T crit time, μ time is set to 1. The membership function of the target residence time uses the Sigmoid function to calculate the membership degree of the intrusion target. Among them, the parameter k is the slope coefficient, which can control the steepness of the S function, that is, the degree of change of μ time with time; the parameter T mid is the midpoint time, that is, the center symmetry point of the S function. At this time position, μ time is 0.5; T<00000​​​​​​​​​​A fuzzy rule base is a formal representation of domain knowledge and experience. Its purpose is to encode and store the knowledge and experience accumulated by human experts in a specific field in the form of "if...then..." rules. The fuzzy rule base is the core component of fuzzy logic algorithms. By storing and managing the rule information required for fuzzy reasoning, it provides the foundation and basis for fuzzy logic reasoning. The design of the fuzzy rule base should meet the following rules:

[0188] (1) The fuzzy rule base should cover all input combination states and ensure that there are no blind spots in decision making;

[0189] (2) Full spatial division of the four characteristic variables of the airport perimeter intrusion targets: target type, target behavior, target distance, and target residence time;

[0190] (3) Use natural language to describe the rules, that is, the IF-THEN structure, which is expressed as follows:

[0191] IF(μ type ∈L type )AND(μ behavior ∈L behavior )AND(μ distance ∈L distance )AND(μ distance

[0192] ∈L time ) THEN Warning Level = C m

[0193] Among them, L variable is the fuzzy level of characteristic variable, C m The ultra-high / high / medium / low warning level output by the rule; L type is the fuzzy level of the characteristic variable of the target type, L behavior is the fuzzy level of the characteristic variable of the target behavior, L distance is the fuzzy level of the characteristic variable of target distance, L time is the fuzzy level of the characteristic variable of target residence time;

[0194] (4) Fuzzy rule strength calculation should be performed in accordance with the AND calculation method to calculate the rule strength information that the intrusion target characteristic variables meet. The expression is:

[0195] μ rulei =min(μ type ,μ behavior ,μ distance ,μ time )×ω i

[0196] Where μ rulei Adapt the activation strength of the i-th rule to the intrusion target, ωi is the rule weight of the i-th rule.

[0197] Once the fuzzy rule base is designed, it is saved to the database and can be optimized and updated based on the actual operation of the airport perimeter to dynamically meet the security needs of the airport perimeter. The fuzzy rule strength calculation formula is combined with the actual perimeter situation to adapt to the actual use of the perimeter.

[0198] S303, fuzzy reasoning and defuzzification of intrusion alarm features;

[0199] The fuzzy reasoning of intrusion alarm features is based on fuzzified feature variables and fuzzy rule base, and takes the intrusion target feature variables as input items. It performs reasoning according to the steps of feature variable fuzzification, rule strength evaluation, rule consequence aggregation and defuzzification, and finally obtains the intrusion target warning level.

[0200] Exemplarily, the defuzzification is a key link in fuzzy logic reasoning, and its function is to aggregate the fuzzy results of multiple rules through logical operations based on the fuzzified characteristic variables (such as the membership value of target type, behavior, distance, and residence time) and the fuzzy rule base, and finally output a single, definite warning level so that the system can execute the corresponding linkage strategy.

[0201] Fuzzification of characteristic variables is a key prerequisite for fuzzy logic reasoning. Its core is to convert precise characteristic data of an intruder (e.g., distance 5.2 meters, residence time 15 seconds) into membership values ​​within a fuzzy set (e.g., "close distance" has a membership of 0.8), providing a foundation for subsequent rule reasoning. This involves determining input characteristic variables, designing membership functions, and performing fuzzy level classification.

[0202] Exemplarily, the rule strength evaluation is to perform a logical operation on each rule antecedent (precondition) according to the fuzzy rule strength calculation formula based on a predefined fuzzy rule base (IF-THEN rule) to obtain the activation strength of the rule consequent (conclusion);

[0203] The rule consequent aggregation combines the fuzzy sets of all activated rule consequents into a comprehensive output fuzzy set, using the maximum method MAX according to arg max(μ rulei ) method, the results of multiple rules are superimposed and merged, μ rule =max(μ rule1 ,μ rule2 ,μ rule3 …μ rulen ) derives the alarm rule strength value, and then dynamically divides it according to the activation strength threshold maintained by the system. Common methods of dynamic optimization include manual configuration, data-driven optimization, and real-time adaptation;

[0204] Manual configuration includes: manually adjusting the threshold range through the management interface (e.g., in sensitive areas of airports, the "high level" threshold range can be narrowed to improve response sensitivity);

[0205] Data-driven optimization includes: automatically fitting optimal thresholds based on historical intrusion event data through machine learning (such as clustering algorithms) to adapt to environmental differences at different airports;

[0206] Real-time adaptation includes dynamically scaling thresholds based on real-time weather and passenger flow data (e.g., increasing false alarm tolerance during heavy rain by temporarily expanding the "low level" range). The alert activation intensity range generally includes four levels: ultra-high [0.8, 1], high [0.6, 0.8], medium [0.3, 0.6], and low [0.0, 0.3].

[0207] S304, the intelligent alarm classification processing center saves the intrusion target warning level obtained by fuzzy algorithm reasoning into the database, and pushes the intrusion target coordinate point and the intrusion target warning level to the linkage control center through the HTTP protocol.

[0208] S4, based on the priority judgment of the pushed intrusion target warning level, uses the OPENCV algorithm library to mark the trajectory of multiple coordinate points closest to the intrusion target in the surrounding panoramic base map, displays the current movement trajectory of the intrusion target, and judges the intrusion target intention.

[0209] Exemplarily, the intrusion target coordinate point is the pixel position of the intrusion target in the panoramic image. When the perimeter scanning device rotates one circle, the position of the intrusion target may change, thereby causing the intrusion target warning level to change.

[0210] For example, the OPENCV algorithm library is an open source algorithm library for computer vision, which has high efficiency and cross-platform characteristics and can perform various processing and operations on images. Compared with the traditional airport perimeter system that can only mark the real-time position of the intrusion target and lacks historical trajectory association and motion trend analysis, the present invention uses multi-frame coordinate association, visualization enhancement, real-time optimization and other methods to visualize the motion trajectory of the intrusion target in real time, assisting airport perimeter staff to judge the intention of the intrusion target.

[0211] The trajectory annotation steps include: 1) obtaining the six most recent valid real-time coordinate points (PositionX, PositionY) of the intrusion target in chronological order to avoid interference from outdated data; 2) reading the dual-band panoramic image (visible light / infrared) through cv2.imread() as the base map for trajectory annotation; 3) using the cv2.line() function to draw line segments point by point for the six cached coordinate points in chronological order; 4) setting the line color according to the target warning level (such as red for ultra-high level and yellow for high level) to enhance visual distinction; 5) triggering the trajectory annotation process based on the latest panoramic image generated by the periscopic device to automatically update the trajectory. In addition, when annotating the trajectory, OpenCV's cv2.GaussianBlur() is used to smooth the coordinate points and filter out noise points caused by device jitter or short-term occlusion of the target. This completes the real-time trajectory tracking of the intrusion target under the panoramic image.

[0212] For example, the linkage control center performs trajectory tracking and video tracking of the intruder target based on the intruder target's coordinate points and the intruder target's warning level. The trajectory tracking algorithm (which connects the intruder target's six nearest coordinate points using the OpenCV algorithm to form the intruder target's trajectory) calculates and connects the intruder target's real-time coordinate points, displaying the intruder target's trajectory in real time on the front-end monitoring system to help monitoring personnel determine the intruder's intentions. The video tracking uses a coordinate mapping algorithm to map the real-time intruder target's coordinate points to the PTZ parameters of the ball camera, and calls the ball camera SDK in real time to adjust the ball camera's angle and zoom ratio according to the intruder target's coordinate position, thereby tracking the intruder target through real-time video.

[0213] For example, the linkage control center can automatically identify and lock the most threatening target based on the warning level of the intrusion target. By mapping the intrusion target coordinate point in the perimeter panoramic base map to the horizontal rotation angle (P), pitch angle (T) and zoom factor (Z) of the dome camera in real time, accurate tracking of the intrusion target is achieved. At the same time, the linkage control center obtains the dome camera information bound to the current panoramic scanning device from the database, and constructs a three-dimensional space coordinate conversion model based on the mapping relationship between the panoramic coordinate system and the physical coordinate system of the dome camera, automatically controlling the dome camera to turn in the direction of the intrusion target for video tracking confirmation.

[0214] The dome camera is a PTZ (Pan-Tilt-Zoom) type dome camera, which is widely used in the field of security monitoring and has horizontal rotation (Pan), vertical pitch (Tilt) and lens zoom (Zoom) functions.

[0215] The initial horizontal angle of the dome camera is 0°, which is consistent with the initial direction setting of the perimeter scanning device. The linkage control center calculates the PTZ parameters of the dome camera based on the coordinates of the intrusion target (PositionX, PositionY), the vertical field of view angle (ForVer) and the pitch angle (DangleVer) of the perimeter scanning device. The specific mapping formula is as follows:

[0216] The horizontal angle (P) uses a polar coordinate mapping algorithm to convert the target pixel coordinates (PositionX / ImgWidth) into the horizontal rotation angle of the dome camera:

[0217]

[0218] Where ImgWidth and ImgHeight are the pixel width and height of the panorama, θ offset The initial angle calibration offset between the scanning device and the ball camera is pre-measured and stored by the device calibration program. When the scanning device and the ball camera are at the same angle θ offset =0;

[0219] Tilt is calculated by combining the vertical field of view angle ForVer and the pitch angle DangleVer of the dome camera using a triangular geometry algorithm:

[0220]

[0221] In the formula, the negative sign indicates the engineering convention that the downward direction of the ball camera is the positive direction;

[0222] This formula essentially converts the vertical offset of a pixel into a pitch angle in physical space.

[0223] The zoom factor Z is:

[0224]

[0225] Where Z min ,Z max They are the minimum and maximum zoom ratios preset for the speed dome camera respectively;

[0226] This formula ensures that the zoom factor is automatically increased when the target is below the panoramic view (close distance) to achieve special monitoring.

[0227] In short, the linkage control center sends PTZ positioning commands to the dome camera based on real-time calculated PTZ parameters. The dome camera's actuator uses a trajectory planning algorithm based on a trapezoidal velocity curve to achieve smooth control of horizontal (pan) and vertical (tilt) motion. Upon receiving the command, the dome camera smoothly moves horizontally and vertically according to the P and T parameters. Upon reaching the target position, it performs zoom processing and automatically adjusts the focus based on the Z parameter, thereby obtaining a clear image of the intruder target.

[0228] It can be seen from the above embodiments that the present invention has the effect of reducing the manpower cost of airport security: through intelligent hierarchical early warning and linkage tracking, the frequency of manual inspections is reduced and the emergency response time is shortened.

[0229] Market potential for equipment upgrades: Traditional airport perimeter systems (such as fiber optics and infrared beamforming) face technical bottlenecks. This invention can serve as a next-generation intelligent transformation solution, applicable to over 200 airports nationwide.

[0230] Existing airport perimeter systems still rely primarily on single technologies (such as electronic fencing and video surveillance). Although some systems incorporate AI detection, these systems rely on fixed thresholds for graded warnings (e.g., "trigger a high alarm if within 5 meters of the perimeter") and are unable to dynamically integrate multi-dimensional features (such as behavioral patterns and duration of stay). This invention, for the first time, integrates fuzzy logic reasoning into perimeter security, enabling dynamic, quantitative assessment of intrusion risk and filling the gap in the technology for multi-feature fusion and adaptive classification.

[0231] Existing systems often rely on human staff to determine the danger level of intruders, resulting in a high incidence of false positives and missed alerts, which overwhelms staff. This invention uses fuzzy logic to combine target type, behavior, distance, and duration to automatically determine high, medium, and low risk, reducing labor costs and avoiding resource waste. Furthermore, intelligent perimeter warning technology based on fuzzy logic reasoning enables precise response to intruders, resolving the industry's pain point of "static rules failing to adapt to dynamic threats."

[0232] Traditional technology assumes that high-precision sensors (such as lidar) and fixed threshold algorithms can meet these requirements, but this ignores the diversity of threats and complexity of real-world scenarios. This invention breaks this bias, demonstrating that an intelligent perimeter warning solution based on fuzzy logic reasoning offers superior adaptability and accuracy.

[0233] Example 2, as Figure 1 The airport perimeter graded warning linkage system based on fuzzy logic reasoning provided by the embodiment of the present invention is mainly composed of four core modules: a dual-band panoramic monitoring subsystem, an AI intelligent analysis center, an intelligent alarm graded processing center, and a linkage control center.

[0234] Another object of the present invention is to provide an airport perimeter graded warning linkage system based on fuzzy logic reasoning, comprising:

[0235] A dual-band panoramic monitoring subsystem is used to scan the airport perimeter to generate infrared and visible light panoramic images, which are then transmitted via the MQTT network protocol.

[0236] The AI ​​intelligent analysis center receives infrared and visible light panoramic image data and uses a multimodal target detection algorithm based on deep learning to detect and identify intruder targets in the perimeter defense areas of the infrared and visible light panoramic images. It also pushes warning information about detected and identified intruder targets, which includes the target coordinates and intrusion characteristics.

[0237] The intelligent alarm classification processing center is used to receive and update the infrared and visible light panoramic images sent in real time as the perimeter panoramic base map. It uses fuzzy logic algorithms to classify the intrusion target warning information sent, and makes priority judgments on the intrusion target warning levels. The priority judgment of the intrusion target warning levels is pushed through the HTTP protocol.

[0238] The linkage control center is used to judge the warning level of the intrusion target after pushing the priority, use the OPENCV algorithm library to mark the trajectory of the multiple coordinate points closest to the intrusion target in the panoramic base map of the perimeter, display the movement trajectory of the current intrusion target, and judge the intention of the intrusion target.

[0239] Exemplarily, the dual-band panoramic monitoring subsystem includes multiple sets of circumferential scanning devices, and the parameter information of each circumferential scanning device is stored in a database. The circumferential scanning device is a monitoring device capable of panoramic scanning, with performance characteristics of a horizontal viewing angle of 360°, a vertical viewing angle of 8.7°, a pitch angle range of -20° to +80°, and a scanning speed of 144° / s (i.e., completing one rotation in 2.5 seconds). Its image detector resolution is 26288x640 in infrared light mode and 22692x512 in visible light mode. Each circumferential scanning device is installed at intervals of 200 meters and fixed together with the dome camera on the cross arm of the pole. The horizontal initial steering angle of the circumferential scanning device is consistent with that of the dome camera, and the pole is installed on the inner side of the boundary wall. The parameter information of the circumferential scanning device includes key data such as device number, IP address, port number, longitude, latitude, visible light panoramic image width, visible light panoramic image height, infrared panoramic image width, infrared panoramic image height, vertical field of view angle, and pitch angle.

[0240] like Figure 2 The dual-band panoramic monitoring subsystem establishes a network connection with the perimeter system service via a network switch. After the weekly scanning device completes a full scan, the system sends a device panoramic image message to the AI ​​Intelligent Analysis Center and the Intelligent Alarm Grading Processing Center via the MQTT protocol and saves the message to the perimeter system database. The MQTT message topic is / img / full / realtime, and the device panoramic image message is transmitted in JSON format.

[0241] For example, after completing the scan for the weekly scanning device numbered 001, the message format is as follows:

[0242] {deviceId:001,imgs:[{channelId:0,width:26288,height:640,forVer:8.7,dangleVer:-20.0,img:byte[]},{c channelId:1,width:22692,height:512,forVer:8.7,dangleVer:-20.0,img:byte[]}],time:2024-03-1510:00:00}

[0243] [deviceId is the number of the circular scanning device; imgs is the panoramic image array of the circular scanning device, including infrared panoramic images and visible light panoramic images, where channelId is the channel number, 0 is the infrared panoramic image channel, 1 is the visible light panoramic image channel, width is the panoramic image width, height is the panoramic image height, forVer is the vertical field of view angle of the circular scanning device, dangleVer is the pitch angle of the circular scanning device, img is the panoramic image information, saved in byte array encoding; time is the message upload time].

[0244] The AI ​​intelligent analysis center subscribes to the MQTT topic ( / img / full / realtime) to receive real-time panoramic data streams from the dual-band panoramic monitoring subsystem. The system uses a multi-threaded mechanism to simultaneously analyze the visible and infrared dual-mode panoramic images. Based on the panoramic information, the system sets up a virtual perimeter wall, with a core defense zone set 5 meters inside and outside the perimeter wall, and a warning defense zone set 20 meters inside and outside the perimeter wall.

[0245] The AI ​​intelligent analysis center, based on a deep learning-based multimodal target detection algorithm, detects and identifies intruders in real-time panoramic images, using an adaptive background modeling algorithm to eliminate static environmental interference. When a valid intruder is triggered, the system executes a multi-dimensional analysis process: first, an event code based on the UUID standard is generated and labeled with spatiotemporal coordinates. A multispectral fusion feature vector (including RGB spectrum distribution, infrared radiation intensity, and morphological profile parameters) is then extracted. The timing analysis engine is simultaneously activated, combining the device's historical image library to reconstruct motion trajectories and match behavioral patterns to establish a three-dimensional spatial motion model of the target. The resulting structured feature dataset includes the following parameters: target type (classification code: 00-Other / 01-Personnel / 02-Vehicle / 03-Animal / 04-UAV), target behavior (state code: 00-Stationary / 01-Wandering / 02-Throwing / 03-Climbing), target distance (the shortest spatial Euclidean distance based on the perimeter wall reference plane, in meters), and target residence time (the duration, in seconds, that a target with the same ID remains within the defense zone).

[0246] Furthermore, after the AI ​​intelligent analysis center completes the extraction of the intrusion target feature information, it calls the WebAPI interface ( / api / analysis) of the intelligent alarm classification processing center through the HTTP protocol to upload the intrusion target feature information to the intelligent alarm classification processing center. The intrusion target feature information data format is transmitted in json string format.

[0247] For example, for the intrusion target feature information extracted from the panoramic image of the 001 weekly scanning device, the message sent is as follows:

[0248] {alarmId:00001,timestamp:1742108033,targets:[{channelId:1,targetId:001,origi nalPositionX:12634,originalPositionY:68,pixelHeight:26,pixelWidth:8,type:{value:01,reliability:80},behavior:{value:01,reliability:90},distance:10,time:11},{channelId:1,targ etId:002,originalPositionX:4355,originalPositionY:398,pixelHeight:142,pixelWidth:54,type:{value:01,reliability:95},behavior:{value:02,reliability:85},distance:5,time:19}]};

[0249] Where alarmId is the alarm number; timestamp is the current timestamp; targets is an array of intrusion targets, which can contain information about one or more intrusion targets. Each intrusion target contains the following: channelId is the channel number, 0 for the infrared panoramic channel and 1 for the visible light panoramic channel; targetId is the intrusion target number; originalPositionX is the intrusion target's x-coordinate pixel value; originalPositionY is the intrusion target's y-coordinate pixel value; pixelHeight is the intrusion target's height pixel value; pixelWidth is the intrusion target's width pixel value; type is the intrusion target type, including its content and confidence level; behavior is the intrusion target's behavior, including its content and confidence level; distance is the distance between the intrusion target and the nearest perimeter wall, in meters; and time is the time the intrusion target remains within the perimeter, in seconds.

[0250] For example, the intelligent alarm hierarchical processing center subscribes to the MQTT topic ( / img / full / realtime) to receive device panoramic image messages uploaded by the dual-band panoramic monitoring subsystem in real time. It then stores the visible light and infrared panoramic images separately on a distributed file server. Fence personnel can select a specific panoramic scanning device through a visual front-end interface to view the updated panoramic image information in real time.

[0251] When the intelligent alarm classification processing center receives the intrusion target feature information sent by the AI ​​intelligent analysis center, it performs logical reasoning on the target features based on the fuzzy logic algorithm and ultimately determines the warning level of the intrusion target. The logical reasoning process is as follows: Figure 3 The specific steps are as follows:

[0252] First, the input variable is fuzzified and the intrusion target characteristics are fuzzified. According to a preferred embodiment of this embodiment, the fuzzy level of the input variable is set by the boundary system. For example, the fuzzy level of the target type (high: ≥0.8, medium: 0.5-0.8, low <0.5) corresponds to the following types: people (1.0) = high, drones (0.8) = high, cars (0.6) = medium, animals (0.3) = low; the fuzzy level of the target behavior ... 0.5) corresponds to the following behaviors: climbing (1.0) = high, throwing (0.9) = high, wandering (0.7) = medium, and standing still (0.4) = low; target distance fuzziness level (near: ≥0.75, medium: 0.3-0.75, far: <0.3), with parameters m = 5 and σ = 5 in the distance formula; target residence time fuzziness level (long: ≥0.7, medium: 0.4-0.7, short: <0.4), with parameters k = 0.3, Tmid = 10, and Tcrit = 20 in the residence time formula. After defining the fuzzy levels of the input variables, the membership degree of each intrusion target feature can be calculated using the membership calculation formula. For example, the membership degree can be calculated for the intrusion target feature information extracted from the panoramic image of the 001 weekly scanning device. Among them, the target 001 feature membership is: target type feature membership μ_type = ω_type×(reliability / 100) = 1.0×(80 / 100) = 0.80; target behavior feature membership μ_behavior = ω_behavior×(reliability / 100) = 0.7×(90 / 100) = 0.63; the target distance is 10 meters, so the target distance membership μ_distance = e^[-(dm)^2 / 2σ^2] = e^[-(10-5)^2 / 2×5^2]≈0.60; the target residence time is 11 seconds, so the target residence time membership μ_time = 1 / (1+exp[-k(t-Tmid)]) = 1 / (1+exp[-0.3(11-10)])≈0.57. Similarly, the calculated feature memberships for target 002 are: target type feature membership μ_type = 1.0 × (95 / 100) = 0.95; target behavior feature membership μ_behavior = 0.9 × (85 / 100) = 0.85; target distance membership μ_distance = 1.0 (5 meters is within the perimeter alarm zone. According to the target distance membership formula, when the distance is less than or equal to 5 meters, the membership is 1.0); target residence time membership μ_time = 1 / (1 + exp[-0.3(19-10)]) ≈ 0.94. Therefore, the final feature membership set for target 001 is (0.80 = high, 0.63 = medium, 0.60 = medium, 0.57 = medium), and the feature membership set for target 002 is (0.95 = high, 0.85 = high, 1.0 = close, 0.94 = long).

[0253] Then, based on perimeter security domain knowledge and expert experience, a fuzzy rule base for intrusion alarms is constructed. This fuzzy rule base should cover all possible input combinations and fully partition the intrusion target feature variables. The rule base design uses an IF-THEN structure. According to another preferred embodiment of this embodiment, a partial rule base matrix is ​​as follows:

[0254] R0001: IF μ_type = HIGH AND μ_behavior = HIGH AND μ_distance = NEAR THEN output level = ULTRA HIGH, weight ω1 = 1.0;

[0255] R0002: IF μ_type = high AND μ_behavior = high AND μ_distance = medium AND μ_time = long THEN output level = high, weight ω2 = 0.9;

[0256] R0003: IF μ_type = high AND μ_distance = near THEN output level = high, weight ω3 = 0.8;

[0257] R0004: IF μ_type = high AND μ_behavior = high AND μ_distance = medium AND μ_time = medium THEN output level = high, weight ω4 = 0.8;

[0258] R0005: IF μ_type = high AND (μ_behavior = medium OR μ_time = medium) AND μ_distance = medium THEN output level = medium, weight ω5 = 0.6;

[0259] R0006: IF μ_type = high AND μ_behavior = medium AND μ_distance = far AND μ_time = short THEN output level = low, weight ω6 = 0.3;

[0260] Furthermore, according to the defined fuzzy rule base and the calculation formula for fuzzy rule activation strength, the feature membership sets of target 001 and target 002 are evaluated respectively. Target 001 adapts to the R0005 rule, so the activation strength of target 001 is μ_rule=max(min(μ_type, μ_behavior, μ_distance), min(μ_type, μ_time,

[0261] μ_distance)) × ω5 = max (min (0.80, 0.63, 0.60), min (0.80, 0.57, 0.60)) × 0.6 = 0.36;

[0262] Target 002 adapts to both rules R0001 and R0003, so the activation strength of target 002 is μ_rule=max(μ_rule1, μ_rule3)=max(min(0.95, 0.85, 1.0, 0.94)×1.0, min(0.95, 1.0)×0.8)=0.85.

[0263] Finally, according to the system-maintained warning level intensity grading rules, the final aggregated activation intensity value of the intrusion target is defuzzified. According to another preferred embodiment of this embodiment, the warning level intensity is graded as ultra-high [0.8, 1], high [0.6, 0.8], medium [0.3, 0.6], and low [0.0, 0.3]. Therefore, the warning level of target 001 is 0.36 (medium), and the warning level of target 002 is 0.85 (ultra-high).

[0264] For example, when the intelligent alarm classification processing center performs fuzzy logic reasoning on the intrusion target alarm information and ultimately derives the final warning level, the linkage control center will link the corresponding dome camera to conduct video surveillance and tracking of the intrusion target with the highest warning level, and at the same time automatically alert the staff according to the perimeter security disposal process. According to the calculation of the intrusion target warning level in this embodiment, target 002 is the target with the highest warning level. At the same time, based on the coordinate information of target 002 (4355, 20) and the dome camera PTZ parameter mapping formula, it is obtained that P = (PositionX / ImgWidth) × 360 = (4355 / 22692) × 360 ≈ 69.1; T = -(DangleVer + arctan([2×|ImgHeight / 2-PositionY|×tan(ForVer / 2)] / ImgHeight)) ≈ 22; Z = ZoomMin + (ZoomMax-ZoomMin) × [(ImgHeight-PositionY) / ImgHeight] 2 = 2 + (20 - 2) × [(512 - 398) / 512] 2 ≈ 3. This gives the dome camera's PTZ parameters as (69.1, 22, 3). The linkage control center sends the PTZ parameter values ​​to the dome camera. The dome camera moves according to these parameters, with a horizontal angle of 69.1°, a vertical angle of 22° (positive when facing downward), and a magnification of 3x.

[0265] As another example, staff can use the front-end alarm monitoring interface to intuitively view real-time visible light and infrared panoramic images from each perimeter scanning device, while also obtaining the warning levels and characteristic information of various intrusion targets reported by the intelligent alarm classification processing center. Furthermore, the alarm monitoring interface can display the movement trajectory of intrusion targets in real time within the panoramic image. It automatically provides a real-time video monitoring page for the intrusion target with the highest warning level. Of course, staff can also manually control the dome camera to observe and monitor other intrusion targets as needed. Through these functions, staff can accurately determine the intentions of intrusion targets and promptly initiate emergency response and security control measures for intrusion alarms, ensuring the safety and stability of the relevant areas.

[0266] To further illustrate the effects of the embodiments of the present invention, the following experiments were conducted.

[0267] In order to verify the analysis and processing effect of the present invention, a real airport fence intrusion scene was simulated and processed using the method of the present invention to verify its ability to identify targets of different dangerousness, the accuracy of warning level judgment, and the effect of linkage ball camera video tracking and trajectory labeling, to implement a simulation program and conduct simulation experiments. First, a simulation experiment was conducted on the method of the present invention in combination with multiple historical infrared and visible light panoramic images of airport fences. Multiple historical infrared and visible light panoramic images of airport fences were collected, covering different time periods, weather conditions and lighting conditions to ensure the diversity of image data. The target information in the image was annotated in detail, including target type (people, cars, animals, drones), target location (accurate to meters, range [0m, 50m]), target behavior (stationary, wandering, climbing, throwing) and target residence time (accurate to seconds, range [0s, 30s]). Furthermore, parameters and rule definitions were set for fuzzy logic reasoning. For example, the target type fuzzy level was set to (high: ≥0.8, medium: 0.5-0.8, low <0.5), the target behavior fuzzy level was set to (high: ≥0.8, medium: 0.5-0.8, low <0.5), the target distance fuzzy level was set to (near: ≥0.75, medium: 0.3-0.75, far: <0.3), the target residence time fuzzy level was set to (long: ≥0.7, medium: 0.4-0.7, short: <0.4), the number of fuzzy rules was set to 64, and the alarm intensity was set to ultra-high [0.8, 1], high [0.6, 0.8], medium [0.3, 0.6], and low [0.0, 0.3]. Based on these fuzzy levels and combined with practical experience and domain knowledge in airport perimeter security, 64 fuzzy rules were constructed. At the same time, the alarm intensity levels are set to ultra-high [0.8, 1], high [0.6, 0.8], medium [0.3, 0.6], and low [0.0, 0.3]. A simulation program is written, and the collected historical image data is used to generate 100 sets of simulation data according to the range requirements of target type, location, behavior, and residence time. The infrared and visible light panoramic images of the perimeter are simulated and pushed to the system of the present invention. The present invention realizes the processing of panoramic images, target feature extraction, fuzzy logic reasoning, warning level judgment, ball camera linkage control and trajectory marking, etc. During the processing, the membership function parameters are continuously optimized and adjusted to achieve the best target recognition effect.

[0268] During this simulation, 11 targets were identified as extremely dangerous, 24 targets were identified as high dangerous, 36 targets were identified as medium dangerous, 15 targets were identified as low dangerous, and 14 targets were not dangerous, so the perimeter alarm was not triggered. By comparing the warning level identified in the simulation experiment with the preset actual danger level, the statistical warning level judgment accuracy was high, the average deviation rate of the linked dome camera video tracking was low, and the trajectory annotation and tracking effect was good. Meeting the actual needs of airport security, compared with traditional perimeter systems, this invention more effectively assists perimeter monitoring personnel in quickly identifying and monitoring high-risk targets.

[0269] The above description is only a preferred specific implementation method of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with this technical field within the technical scope disclosed by the present invention and within the spirit and principles of the present invention should be covered by the scope of protection of the present invention.

Claims

1. A method for linking airport perimeter graded warnings based on fuzzy logic reasoning, characterized in that: The method comprises the following steps: S1: Panoramic scanning of the airport perimeter to generate infrared and visible light panoramic images, which are then sent to the outside world via the MQTT network protocol. S2 receives infrared and visible light panoramic image data, and detects and identifies intruder targets in the perimeter defense area in the infrared and visible light panoramic images using a multimodal target detection algorithm based on deep learning. The detected and identified intruder target warning information is pushed, and the intruder target warning information includes the intruder target coordinates and intrusion feature information. S3 receives and updates the infrared and visible light panoramic images sent in real time as the perimeter panoramic base map. It uses fuzzy logic algorithms to classify the sent intrusion target warning information into different levels, and then prioritizes the intrusion target warning levels. The prioritized intrusion target warning levels are then pushed to the server via HTTP. S4, based on the priority judgment of the pushed intrusion target warning level, uses the OPENCV algorithm library to mark the trajectory of multiple coordinate points closest to the intrusion target in the surrounding panoramic base map, displays the current movement trajectory of the intrusion target, and judges the intrusion target intention.

2. The airport perimeter graded warning linkage method based on fuzzy logic reasoning according to claim 1 is characterized in that: In step S1, the dual-band panoramic monitoring subsystem performs a panoramic scan of the airport perimeter to generate infrared and visible light panoramic images. The dual-band panoramic monitoring subsystem includes multiple panoramic scanning monitoring devices, each of which is installed at intervals and simultaneously mounted with the dome camera on the cross arm of a pole mounted inside the airport perimeter. The pole is installed. By continuously rotating 360 degrees horizontally, each device generates a 360-degree panoramic image at fixed time intervals. The specific steps are as follows: Step 1: Establish a model for the weakening mechanism of interference light disturbing the grayscale image and determine the optimization framework; Step 2: constructing data items for modeling light disturbance; Step 3: Select image prior terms and combine them with data terms to construct a non-optimal feedback model for grayscale image disturbance caused by interference light. Step 4: Use iterative reweighted second-order iterative optimization method to obtain a clear panoramic image.

3. The airport perimeter graded warning linkage method based on fuzzy logic reasoning according to claim 2 is characterized in that: In step 1, a model for the weakening mechanism of interfering light perturbing the grayscale image is established, and an optimization framework is determined, including: (1) Establish a nonlinear weakening model for the grayscale image disturbed by interference light, and the expression is: Where, is the vector of the blurred image that has not been polluted by grayscale, D is the known blur kernel, u is the vector of the original clear image, f is the vector of the final weakened image, and W is the process of the image being polluted by the disturbing light; (2) According to the weakening model, based on the maximum a posteriori probability framework, data fidelity and image prior are combined into a unified image restoration framework. The non-optimal feedback problem corresponds to an optimization problem, which is expressed as: Where, is the image vector in the unified image restoration framework, C(u|f) is the posterior probability; According to the Bayesian formula, we get: C(u|f)∝C(f|u)C(u) Where C(f|u) is the likelihood probability and C(u) is the prior probability; Combining the above two formulas, the equivalent transformation is: Where, is the image vector after equivalent transformation; The following optimization framework is obtained: In the formula, ψ(Du-f) is the data item to be constructed, is the image prior item to be designed; In step 2, the data items for modeling the light disturbance are constructed, including: (1) Select the following nonlinear function to eliminate the influence of disturbance grayscale: Where h and l are function parameters, h controls the nonlinearity of the function at the cutoff point, l controls the cutoff point, D(u) is the blur kernel after removing the influence of the disturbance grayscale, and e is the Reynolds number. (2) Combine the nonlinear function of L1 norm and construct the following data items: Where i is the image pixel index, f is the vector of the final weakened image, S(u) is the data item of the image vector, and Q is the nonlinear function; In step 3, a non-optimal feedback model for grayscale image disturbance caused by interference light is constructed by combining the data items, including: The expression of the prior term constructed based on the image gradient sparse prior is: Where, L w is the sparse prior value of the image gradient of the w-th image node, is the image gradient sparse prior non-optimal feedback, L v is the sparse prior value of the image gradient of the vth image node; In step 4, the iterative reweighted second-order iterative optimization method includes: The non-optimal feedback model is transformed into a weighted second-order iterative optimization method as shown below: Where λ u is the regularization term weight, For weighted second-order iterative optimization, A w is the image gradient sparse prior value weight diagonal matrix in the w-th image node deformation, D w is the blur kernel in the deformation of the w-th image node, A v is the diagonal matrix of image gradient sparse prior weights in the v-th image node deformation, D v is the blur kernel in the deformation of the vth image node.

4. The airport perimeter graded warning linkage method based on fuzzy logic reasoning according to claim 1 is characterized in that: In step S2, receiving infrared and visible light panoramic image data includes: The AI ​​intelligent analysis center subscribes to MQTT events to receive infrared and visible light panoramic images sent by the dual-band panoramic monitoring subsystem. Using a multimodal target detection algorithm based on deep learning, it detects and identifies intruder targets within the perimeter defense areas in the infrared and visible light panoramic images, and pushes warning information about detected and identified intruder targets to the intelligent alarm classification processing center. Intrusion target detection includes target type, target behavior, target distance, and target residence time. Target type includes characteristics of people, vehicles, animals, and drones; target behavior includes characteristics of stationary, wandering, climbing, and throwing; target distance is the vertical distance between the target and the virtual perimeter in the panoramic image; the closer the distance, the higher the risk of the intrusion target; and target residence time is the time the intrusion target remains within the perimeter defense zone. The deep learning-based multimodal target detection algorithm includes: first generating event codes based on the UUID standard and attaching spatiotemporal coordinate tags; then extracting multispectral fusion feature vectors, including RGB spectrum distribution, infrared radiation intensity, and morphological profile parameters; synchronously activating the timing analysis engine, combining the device's historical image library to reconstruct motion trajectories and match behavior patterns, and establish a three-dimensional spatial motion situation model of the target; and finally outputting structured feature dataset parameters, including target type, target behavior, target distance, and target residence time. Establishing a three-dimensional target motion situation model includes: Where X, T, K, and t are target type motion values, target behavior motion trend motion values, target distance, and target retention time, respectively; G is the target motion matching parameter; J is the multispectral fusion feature sparseness; z is the vertical target motion area; F is the multispectral fusion feature vector; x is the horizontal target motion area; y is the vertical target motion area; and H is the target motion area. z is the vertical target trending motion coefficient, and I is the target motion trajectory radius.

5. The airport perimeter graded warning linkage method based on fuzzy logic reasoning according to claim 1 is characterized in that: In step S3, the fuzzy logic algorithm is used to classify the intrusion target warning levels and to determine the priority of the intrusion target warning levels, including: S301, fuzzy design of invasion target characteristic variables; S302, design of intrusion alarm fuzzy rule base; S303, fuzzy reasoning and defuzzification of intrusion alarm features; S304, the intelligent alarm classification processing center saves the intrusion target warning level obtained by fuzzy algorithm reasoning into the database, and pushes the intrusion target coordinate point and the intrusion target warning level to the linkage control center through the HTTP protocol.

6. The airport perimeter graded warning linkage method based on fuzzy logic reasoning according to claim 5 is characterized in that: In step S301, the fuzzy design of the intrusion target characteristic variables includes: The target type membership function expression is: Where μ type is the target type membership, ranging from [0,1]; ω type is the target type risk weight, ranging from [0,1]; reliability is the target behavior confidence; The target behavior membership function expression is: Where μ behavior is the target behavior membership, ranging from [0,1]; ω behavior is the risk weight of the target behavior, ranging from [0,1]; The target distance membership function expression is: Where μ distance is the target distance membership, ranging from [0, 1]; d is the vertical distance between the target and the boundary wall, m is the center point, m = 5; σ is the standard deviation, σ = 5; The target residence time membership function expression is: Where μ time is the target residence time membership, ranging from [0,1]; k is the slope coefficient, T mid is the midpoint time, T crit is the critical time; In step S302, the intrusion alarm fuzzy rule base is designed, including: The fuzzy rule base is designed to meet the following rules: (1) The fuzzy rule base should cover all input combination states; (2) Full spatial division of the four characteristic variables of the airport perimeter intrusion targets: target type, target behavior, target distance, and target residence time; (3) Use natural language to describe the rules, that is, the IF-THEN structure, which is expressed as follows: IF(μ type ∈L type )AND(μ behavior ∈L behavior )AND(μ distance ∈L distance )AND(μ distance ∈L time ) THEN Warning Level = C m Among them, L variable is the fuzzy level of characteristic variable, C m The ultra-high / high / medium / low warning level output by the rule; L type is the fuzzy level of the characteristic variable of the target type, L behavior is the fuzzy level of the characteristic variable of the target behavior, L distance is the fuzzy level of the characteristic variable of target distance, L time is the fuzzy level of the characteristic variable of target residence time; (4) Fuzzy rule strength calculation should be performed in accordance with the AND calculation method to calculate the rule strength information that the intrusion target characteristic variables meet. The expression is: m rulei =min(μ type ,m behavior ,m distance ,m time )×ω i Where μ rulei Adapt the activation strength of the i-th rule to the intrusion target, ω i is the rule weight of the i-th rule.

7. The airport perimeter graded warning linkage method based on fuzzy logic reasoning according to claim 4 is characterized in that: In step S303, fuzzy reasoning and defuzzification of intrusion alarm features include: Based on fuzzy feature variables and fuzzy rule base, and taking the intrusion target feature variables as input, the method performs reasoning according to the steps of feature variable fuzzification, rule strength evaluation, rule consequence aggregation and defuzzification, and finally obtains the intrusion target warning level. The rule strength evaluation is to perform a logical operation on each rule antecedent according to the fuzzy rule strength calculation formula based on the predefined fuzzy rule base to obtain the activation strength of the rule consequent; The rule consequent aggregation combines the fuzzy sets of all activated rule consequents into a comprehensive output fuzzy set, using the maximum value method MAX according to argmax(μ rulei ) method, the results of multiple rules are superimposed and merged, μ rul e=max(μ rule1 ,μ rule2 ,μ rule3 …μ rulen ) derives the alarm rule strength value, and then dynamically divides it according to the activation strength threshold maintained by the system. Common methods of dynamic optimization include manual configuration, data-driven optimization, and real-time adaptation; Manual configuration involves manually adjusting the threshold range through the management interface; data-driven optimization involves automatically fitting the optimal threshold through machine learning based on historical intrusion event data to adapt to the environmental differences of different airports; Real-time adaptation includes: combining real-time weather and passenger flow data to dynamically scale the threshold range; the warning level activation intensity range includes four levels: ultra-high [0.8, 1], high [0.6, 0.8], medium [0.3, 0.6], and low [0.0, 0.3].

8. The airport perimeter graded warning linkage method based on fuzzy logic reasoning according to claim 1 is characterized in that: In step S4, the OPENCV algorithm library is used to mark the trajectories of the multiple coordinate points closest to the intrusion target in the surrounding panoramic base map, display the current movement trajectory of the intrusion target, and judge the intrusion target intention, including: The linkage control center can automatically identify and lock the most threatening target based on the warning level of the intrusion target; it can track the intrusion target by mapping the intrusion target coordinate point in the panoramic base map of the perimeter to the horizontal rotation angle P, pitch angle T and zoom factor Z of the ball camera in real time; at the same time, the linkage control center obtains the ball camera information bound to the current panoramic scanning device from the database, and builds a three-dimensional space coordinate conversion model based on the mapping relationship between the coordinate system of the panoramic base map of the perimeter and the physical coordinate system of the ball camera, automatically controlling the ball camera to turn towards the intrusion target direction for video tracking confirmation.

9. The airport perimeter graded warning linkage method based on fuzzy logic reasoning according to claim 8 is characterized in that: The linkage control center calculates the PTZ parameters of the ball camera based on the coordinates of the intrusion target (PositionX, PositionY), the vertical field of view angle ForVer, and the pitch angle DangleVer of the periscope device. The specific mapping process is as follows: The horizontal angle P uses a polar coordinate mapping algorithm to convert the target pixel coordinates (PositionX / ImgWidth) into the horizontal rotation angle of the dome camera: Where ImgWidth and ImgHeight are the pixel width and height of the panorama, θ offst The initial angle calibration offset between the scanning device and the ball camera is pre-measured and stored by the device calibration program. When the scanning device and the ball camera are at the same angle θ offst =0; The vertical pitch T is calculated by combining the vertical field of view angle ForVer and the pitch angle DangleVer of the dome camera using a triangular geometry algorithm: In the formula, the negative sign indicates the engineering convention that the downward direction of the ball camera is the positive direction; The zoom factor Z is: Where Z min ,Z max They are respectively the minimum and maximum zoom ratios preset for the speed dome camera.

10. An airport perimeter graded warning linkage system based on fuzzy logic reasoning, characterized by: The system implements the airport perimeter graded warning linkage method based on fuzzy logic reasoning as described in any one of claims 1 to 9, and the system includes: A dual-band panoramic monitoring subsystem is used to scan the airport perimeter to generate infrared and visible light panoramic images, which are then transmitted via the MQTT network protocol. The AI ​​intelligent analysis center receives infrared and visible light panoramic image data and uses a multimodal target detection algorithm based on deep learning to detect and identify intruder targets in the perimeter defense areas of the infrared and visible light panoramic images. It also pushes warning information about detected and identified intruder targets, which includes the target coordinates and intrusion characteristics. The intelligent alarm classification processing center is used to receive and update the infrared and visible light panoramic images sent in real time as the perimeter panoramic base map. It uses fuzzy logic algorithms to classify the intrusion target warning information sent, and makes priority judgments on the intrusion target warning levels. The priority judgment of the intrusion target warning levels is pushed through the HTTP protocol. The linkage control center is used to judge the warning level of the intrusion target after pushing the priority, use the OPENCV algorithm library to mark the trajectory of the multiple coordinate points closest to the intrusion target in the panoramic base map of the perimeter, display the movement trajectory of the current intrusion target, and judge the intention of the intrusion target.

Images