Unmanned aerial vehicle detection and control method and system for regional defense

Abstract

The invention relates to an unmanned aerial vehicle detection and control method and system for regional defense, and the method comprises the steps: dividing a to-be-defended region, and setting a differentiated flight forbidding threshold and a detection response priority; secondly, constructing a multi-dimensional cooperative detection network of radio spectrum monitoring, millimeter wave radar, photoelectric tracking and an acoustic sensor based on sub-region division, and collecting multi-source data of the unmanned aerial vehicle in real time; judging a sub-region to which the unmanned aerial vehicle belongs through coordinate matching, screening out over-limit target data, and extracting a target feature set of flight speed, height, communication protocol type, radar reflection cross section and voiceprint information; and inputting a target dynamic classification model, and dynamically outputting three threat levels of a harmless civil type threat level, a suspicious reconnaissance type threat level and a high-risk aggressive type threat level. The hierarchical management and control are executed in combination with the sub-region priority and the threat level, the management and control effect is dynamically evaluated based on the preset evaluation index, and the detection parameter and the management and control strategy are adaptively adjusted, so that the intrusion risk of the core region is effectively reduced, and the intellectualization and adaptive ability of regional defense are improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of unmanned aerial vehicle management and control, and particularly relates to an unmanned aerial vehicle detection and management and control method and system for regional defense. BACKGROUND

[0002] With the rapid development of unmanned aerial vehicle technology, the number and application scenarios of consumer and industrial unmanned aerial vehicles have increased significantly, which has brought convenience but also serious safety hazards. According to industry statistics, there were more than 5,000 unmanned aerial vehicle-related safety incidents worldwide in 2023, of which unauthorized intrusion into core sensitive areas accounted for 32%, and some incidents even involved malicious intrusion with dangerous goods, posing a direct threat to regional safety. Against this background, how to achieve efficient detection and accurate control of intruding unmanned aerial vehicles has become a core problem that needs to be solved in the security field. Current mainstream unmanned aerial vehicle detection technologies mainly rely on single or limited dimension detection devices, which cannot meet the needs of complex regional defense. Unmanned aerial vehicle detection technologies have the following limitations: Single detection dimension, high miss and false detection rate: Existing technologies mostly use single radar detection or radio spectrum monitoring, such as capturing unmanned aerial vehicle position information only through millimeter wave radar or identifying communication signals only through radio modules. However, low, slow, and small unmanned aerial vehicles (such as those weighing <2 kg and flying at a height of <100 m) have a very small radar cross section (RCS) (usually ≤0.01 m2), which is easily blocked by buildings, trees, and other obstacles, leading to radar miss detection. At the same time, the radio signals of civilian devices including mobile phones and wireless cameras are easily confused with unmanned aerial vehicle communication signals, resulting in a false detection rate of radio spectrum monitoring of more than 15%, making it difficult to accurately filter out real unmanned aerial vehicle targets. No regional differentiated detection strategy, unreasonable resource allocation: Existing detection systems mostly use a global uniform deployment mode, without dividing sub-regions according to the security levels of the defense regions, resulting in insufficient density of detection devices in core sensitive areas and redundant devices in peripheral non-sensitive areas, causing resource waste. For example, an airport defense system uses the same density of radar deployment around the runway and around the parking lot, which cannot meet the high sensitivity detection needs of the runway area (requiring positioning accuracy ≤0.5 m) and results in less than 30% utilization rate of devices in the parking lot area. Weak multi-source data fusion capability, low data value utilization rate: Some detection systems integrate multiple types of devices such as radar, photoelectric, and acoustic devices, but lack effective data fusion mechanisms: radar data and photoelectric data are not synchronized in time stamp, and acoustic data is used directly without filtering environmental noise, resulting in the inability of multi-source data to form synergy and the difficulty of extracting accurate unmanned aerial vehicle features (such as flight speed change rate and voiceprint spectrum peak value), affecting the accuracy of subsequent threat determination. The existing unmanned aerial vehicle threat grading technology has the following deficiencies: threat grading is a prerequisite for precise control, but the existing technology has obvious defects in grading logic and model adaptability: The grading standard is single and does not combine regional security levels: The existing method mainly determines the threat level according to the flight speed of the unmanned aerial vehicle, whether it carries a camera and other simple parameters, and does not consider the safety priority of the region where the unmanned aerial vehicle is located. For example, the same civil aerial unmanned aerial vehicle can be determined as harmless in a general park area, but should be determined as suspicious in a sensitive core area. The existing system does not realize such differentiated grading, which can easily lead to misjudgment of threats in core areas or excessive control in peripheral areas. In the control execution link, the existing technology is difficult to realize graded control and resource optimization, and the defects of the existing unmanned aerial vehicle control technology include: The control measures do not match the threat level, and the response efficiency is low: The existing system mainly adopts a one-size-fits-all control method, such as starting electromagnetic interference for all invading unmanned aerial vehicles, without differentiated execution according to threat levels. For harmless civil unmanned aerial vehicles, excessive interference will affect normal communication; for high-risk attack unmanned aerial vehicles, if only a single interference measure is adopted (without combining physical interception), it is difficult to effectively prevent intrusion, and the control response time of the existing system is generally ≥5s, which cannot meet the emergency demand of ≤1s in core areas. The cross-regional coordination capability is weak, and the multi-target intrusion response is insufficient: When multiple unmanned aerial vehicles simultaneously invade different sub-regions, the existing system lacks a unified resource scheduling mechanism, and each regional control device works independently, which can easily lead to insufficient resources in core areas and idle resources in peripheral areas, and the success rate of multi-target capture is less than 70%. At the same time, the cross-regional data sharing delay is high. Therefore, it is necessary to improve the existing unmanned aerial vehicle detection and control process to solve the technical problems existing in the above-mentioned existing technology. SUMMARY

[0003] The purpose of the present application is to provide an unmanned aerial vehicle detection and control method and system for regional defense, which solves the technical problems existing in the above-mentioned existing technology.

[0004] To solve the above technical problems, the technical solution adopted by the present application is as follows: The first aspect provides an unmanned aerial vehicle detection and control method for regional defense, comprising the following steps: S1: dividing the region to be defended into at least one core defense sub-region and at least two peripheral warning sub-regions, setting the unmanned aerial vehicle no-fly threshold and detection response priority of each sub-region; S2: Constructing a multi-dimensional cooperative detection network based on sub-region division: Deploying detection devices in each sub-region, and collecting multi-source data of each target UAV in the corresponding spatial range by the detection devices in each sub-region; S3: Judging the sub-region where the UAV is located and whether it exceeds the no-fly threshold of the sub-region based on multi-source data, obtaining target data exceeding the threshold, extracting features from the target data, and obtaining a target feature set including flight speed, height, communication protocol type, radar cross section, and voiceprint information for each sub-region; S4: Inputting the extracted feature set into a target dynamic classification model, and outputting the threat level of each target UAV in real time based on the feature set by the target dynamic classification model; S5: Executing a hierarchical dynamic control strategy according to the detection response priority of each sub-region and the corresponding threat level of the target UAV; S6: Continuously collecting state parameters of each target UAV after executing the hierarchical dynamic control strategy, dynamically evaluating the control effect based on the detection response priority of each sub-region and the preset evaluation index, and adaptively adjusting the detection device parameters and control strategy of each sub-region according to the evaluation result.

[0005] Preferably, the specific process of step S1 is as follows: S11: Spatially dividing the region to be defended to determine at least one core defense sub-region and at least two peripheral warning sub-regions; S12: Setting a no-fly threshold for each sub-region; S13: Setting the detection response priority of each sub-region, with the core defense sub-region having the highest priority, the inner peripheral sub-region having a medium priority, and the outer peripheral sub-region having a basic priority.

[0006] Preferably, the specific process of step S2 is as follows: S21: Based on each sub-region and its corresponding detection priority, deploying detection devices in each sub-region to construct a multi-dimensional cooperative detection network, including a radio spectrum monitoring module, a millimeter wave radar array, an optoelectronic tracking device, and an acoustic sensor; S22: The radio spectrum monitoring module uses frequency hopping scanning technology to capture real-time radio spectrum monitoring data including UAV communication band signals; The millimeter wave radar array realizes dynamic tracking of low, slow, and small targets through an adaptive beamforming algorithm, and collects millimeter wave radar data; The optoelectronic tracking device obtains optoelectronic data containing UAV visual positioning information; The acoustic sensor identifies the propeller noise of the UAV based on a voiceprint feature library to obtain acoustic sensor data.

[0007] Preferably, the specific process of step S3 is as follows: S31: Coordinate calibration is performed by matching the coordinates of the unmanned aerial vehicle with the sub-region boundary coordinates preset in the defense area GIS geographic information system; S32: The calibrated real-time coordinates of the unmanned aerial vehicle are matched with the sub-region boundaries; S33: The unmanned aerial vehicle data with the added sub-region attribution label is compared in real time from three dimensions of height, speed, and stay duration. If any dimension exceeds the standard, it is marked as an out-of-limit target, and its original data is included in the out-of-limit target data set. If the data does not exceed the standard, it is marked as a compliant target, and only the basic record is retained, without entering the subsequent feature extraction link; S34: The out-of-limit target data set is preprocessed to eliminate interference data and perform data space-time alignment; S35: Based on the preprocessed out-of-limit target data set, corresponding feature extraction is performed on each type of data to generate a target feature set for each sub-region.

[0008] Preferably, the corresponding feature extraction process for each type of data in step S35 is as follows: For millimeter wave radar data, instantaneous speed is calculated, and flight speed features including mean and variance of speed are extracted. RCS value is calculated based on radar echo power, and radar cross section (RCS) features including mean and fluctuation range of RCS are extracted; For millimeter wave radar data and photoelectric data, radar height H1 and photoelectric height H2 are obtained, and flight height features including maximum height and duration of exceeding the standard are extracted; For radio spectrum monitoring data, communication protocol type features are extracted by analyzing signal modulation method, frame structure, and matching protocol library; For acoustic sensor data, frequency spectrum peak value, bandwidth, and period of noise signal are extracted, and acoustic fingerprint information features are obtained by matching with the unmanned aerial vehicle acoustic fingerprint library.

[0009] Preferably, the target dynamic classification model is based on traditional CNN, with the addition of a sub-region feature attention module and a multi-dimensional feature fusion layer. The architecture is divided into 3 layers: Convolution layer: 3 convolution kernels are used to extract features from the standardized input vector, capture local feature associations, and output feature maps; Attention layer: Feature weights are dynamically assigned based on sub-region attribution labels. The core area feature set is given 2 times the weight for communication protocol encryption level and acoustic fingerprint abnormality, and the peripheral area is given 1.5 times the weight for speed exceeding the standard amplitude and stay duration exceeding the standard, to ensure that high-priority area key features are identified first; Fully connected layer: The feature maps output by the convolution layer and the attention layer are flattened into 128-dimensional vectors, and non-linear transformation is performed through 2 fully connected layers to finally output a 3-dimensional vector corresponding to the probability values of the 3 threat levels.

[0010] Preferably, the specific process of step S4 is as follows: S41: Perform structural conversion on the target feature set to convert it into a numerical matrix format that can be recognized by the target dynamic classification model: S42: Normalize the input vector using the Z-Score standardization algorithm; S43: Take a target feature set in the core area as an example. The convolution kernel calculates the local feature response through the sliding window, and the model generates an attention weight matrix through the Sigmoid function. In the model training stage, the error between the predicted value and the true label is calculated through the cross-entropy loss function, and the Adam optimizer is used to adjust the parameters of each layer in reverse; S44: Threat level probability calculation: The 3-dimensional vector output by the fully connected layer is converted into probability values for each threat level through the Softmax function, and the probability sum is 1: S45: Combine the probability maximum priority rule with the sub-region differentiation threshold value to determine the threat level, and output the final threat level: Preferably, the specific process of step S5 is as follows: S51: Set the correspondence between sub-region priority and resource allocation, and set the priority-threat level control strategy decision matrix: S52: Perform differentiated operations according to the priority-threat level matching: For harmless civilian drones, perform low-intensity control; for suspicious reconnaissance drones, perform medium-intensity control; for high-risk attack drones, perform high-intensity control.

[0011] Preferably, the specific process of step S6 is as follows: S61: Set differentiated evaluation indicators for different control measures: Early warning indicators: harmless civilian drone departure rate, command reception success rate; Interference indicators: suspicious reconnaissance drone image transmission interruption rate, false interference rate; Interception indicators: high-risk attack drone capture success rate, interception response time; S62: According to the evaluation results, adjust the control strategy and resource allocation in reverse: If the core area high-risk attack drone interception success rate is less than 99%, increase the number of intercepted drones or increase the capture net deployment diameter; If the inner peripheral area suspicious reconnaissance drone false interference rate is higher than 0.1%, optimize the electromagnetic interference waveform generation algorithm and increase the identification and filtering of civilian device signals; If the outer peripheral area harmless civilian drone departure rate is less than 90%, upgrade the early warning method and expand the early warning range; S63: Set up an emergency plan update mechanism: The priority-threat level decision matrix is updated monthly based on historical management data: If a new type of UAV is found to be easily misjudged as harmless in the core area, the RCS feature weight of this type of UAV is increased, and the high-risk attack threshold is lowered; If multiple targets cooperatively invade the outer peripheral area multiple times, a multi-target linkage plan for the outer peripheral area is added, a resource allocation path in the adjacent area is preset, and the response time is shortened.

[0012] In a second aspect, a UAV detection and management system for regional defense is used to implement the UAV detection and management method for regional defense, and includes a region division and parameter setting module, a multi-dimensional cooperative detection network module, a target data processing and feature extraction module, a target dynamic classification model, a hierarchical dynamic management and control execution module, and a management and control effect evaluation and strategy optimization module. The modules interact and work cooperatively through a 5G private network or a fiber network. The region division and parameter setting module is used to divide the space of the region to be defended and preset parameters. The multi-dimensional cooperative detection network module is based on the output of the region division and parameter setting module, and different detection devices are arranged in each sub-region to collect multi-source data of target UAVs in real time. The target data processing and feature extraction module is used to receive multi-source data transmitted by the multi-dimensional cooperative detection network module, and perform data filtering, preprocessing and feature extraction. The target dynamic classification model is used to receive target features from the target data processing and feature extraction module, and determine the threat level of the UAV. The hierarchical dynamic management and control execution module is used to perform differential management and control operations according to the response priority of the region division and parameter setting module and the threat level of the target dynamic classification module. The management and control effect evaluation and strategy optimization module is used to continuously collect post-management data from the hierarchical dynamic management and control execution module, and implement effect evaluation and strategy adjustment. The beneficial effects of the present application include: 1. The region to be defended is divided into a core defense sub-region and a peripheral warning sub-region, and differential no-fly thresholds and detection response priorities are set, which avoids the resource waste caused by the uniform arrangement of the prior art. The device utilization rate is improved, which not only ensures high security protection for the core sensitive area, but also avoids device redundancy in the peripheral non-sensitive area. Through sub-region division, a three-level defense line of core defense-inner warning-outer warning is constructed, and relay management is started for UAVs moving across sub-regions (outer warning data is synchronized to the inner layer, and the inner layer is preheated), which solves the problem of weak cross-region cooperation in the prior art. A three-dimensional defense system of outer warning, inner interception and core backup is formed.

[0013] 2. Construct a multi-dimensional cooperative detection network of radio spectrum monitoring + millimeter wave radar + photoelectric tracking + acoustic sensor, realize deep fusion of multi-source data through coordinate calibration and space-time alignment, and form a detection coverage with no dead angle and no blind area. Compared with existing single detection technology, the miss detection rate and false detection rate are reduced.

[0014] 3. The target dynamic classification model gives 2 times weight to the communication protocol encryption level and voiceprint abnormality of the core area target, and gives 1.5 times weight to the speed exceeding standard amplitude and stay time of the peripheral area target, solving the misjudgment problem caused by the fact that the existing technology does not combine the region security level in grading. The core area threat level misjudgment rate is reduced, which avoids the excessive control caused by misjudging harmless in the core area as high risk, and prevents the security loophole caused by misjudging high risk as harmless. 4. Through the priority-threat level decision matrix, harmless civil type is executed with low intensity warning, suspicious reconnaissance type is executed with medium intensity interference, and high risk attack type is executed with high intensity interception, solving the problem of one-size-fits-all control in the existing technology. The core area control response time is strictly controlled within 1s, and the core area high bandwidth data transmission and the core area high power interference are cooperated with multiple interception unmanned aerial vehicles, solving the slow response problem in the existing technology. For core area malicious intrusion events, the control success rate is improved, providing a second-level response security guarantee for core sensitive areas, and effectively preventing the attack of high-risk unmanned aerial vehicles.

[0015] 5. Precise indicators are set for different sub-regions and control types, solving the problem of no standard in the existing technology. According to the evaluation results, the parameters are adjusted reversely, and the emergency plan is updated every month, solving the passive state of execution-failure-repair in the existing technology. The system excavates intrusion rules through historical data backtracking, optimizes equipment layout and control strategy, improves the overall defense efficiency every month, maintains adaptability to complex intrusion scenes in the long term, avoids the decline of defense capability caused by environmental changes or technology iteration, and forms a closed loop of detection-classification-control-evaluation-optimization. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 It is a flowchart of the unmanned aerial vehicle detection and control method for regional defense of the present application.

[0017] Figure 2 It is a defense region division schematic diagram of the present application.

[0018] Figure 3 It is a principle schematic diagram of the target dynamic classification model of the present application. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings Figures 1-3 The present application will be further described in detail: Example 1 Reference is made to the drawings Figure 1The unmanned aerial vehicle detection and management method for regional defense shown includes the following steps: S1: divide the region to be defended into at least one core defense sub-region and at least two peripheral early warning sub-regions, set the unmanned aerial vehicle no-fly threshold and detection response priority of each sub-region, and the unmanned aerial vehicle no-fly threshold includes flight height, speed, and stay time.

[0020] S2: construct a multi-dimensional cooperative detection network based on the sub-region division: different detection devices are arranged in each sub-region, and the detection devices in each sub-region collect multi-source data of each target unmanned aerial vehicle in the corresponding spatial range in real time.

[0021] S3: determine the sub-region where the unmanned aerial vehicle is located and whether it exceeds the no-fly threshold of the sub-region based on the multi-source data, obtain target data that exceeds the threshold, extract features from the target data, and obtain a target feature set including flight speed, height, communication protocol type, radar cross section, and voiceprint information for each sub-region.

[0022] S4: input the extracted feature set into a target dynamic classification model, which outputs the threat level of each target unmanned aerial vehicle in real time based on the feature set, including harmless civilian type, suspicious reconnaissance type, and high-risk attack type, and the threat level of the target unmanned aerial vehicle that exceeds the threshold in the core defense sub-region is automatically upgraded by one level.

[0023] S5: execute a hierarchical dynamic management strategy according to the detection response priority of each sub-region and the corresponding target unmanned aerial vehicle threat level.

[0024] S6: continuously collect the state parameters of each target unmanned aerial vehicle after executing the hierarchical dynamic management strategy, dynamically evaluate the management effect based on the detection response priority of each sub-region and the preset evaluation index, and adaptively adjust the detection device parameters and management strategy of each sub-region according to the evaluation result.

[0025] In this embodiment, the specific process of step S1 is as follows: S11: refer to Figure 2 , spatially divide the region to be defended to determine at least one core defense sub-region and at least two peripheral early warning sub-regions, the two peripheral early warning sub-regions are respectively an inner peripheral sub-region within 1-3 km of the periphery of the core defense sub-region, and an outer peripheral sub-region within 3-10 km of the periphery of the core defense sub-region; S12: set the unmanned aerial vehicle no-fly threshold for each sub-region, wherein the core defense sub-region no-fly threshold is: flight height ≤ 0 m (comprehensive no-fly), speed limit no (once entered, trigger the highest level of response), and stay time ≤ 0 s (prohibit stay); The peripheral early warning sub-region no-fly threshold is: Inner layer peripheral sub-region flight height ≤ 50m, speed ≤ 30km / h, stay duration ≤ 5min, outer layer peripheral sub-region flight height ≤ 120m, speed ≤ 60km / h, stay duration ≤ 15min; S13: Set the detection response priority of each sub-region, the core defense sub-region priority is the highest, the response delay ≤ 10s, the inner layer peripheral sub-region priority is the medium level, the response delay ≤ 2s, and the outer layer peripheral sub-region priority is the basic level, the response delay ≤ 3s.

[0026] The specific process of step S2 is as follows: S21: Based on each sub-region and its corresponding detection priority, the detection equipment is differentiated in each sub-region, and a multi-dimensional cooperative detection network is constructed; In the core defense sub-region, the encryption arrangement density is 2-3 wireless radio spectrum monitoring modules per square kilometer, 1-2 groups of millimeter wave radar arrays per square kilometer, light and electrical tracking equipment covering no dead angle, and acoustic sensors; In the peripheral early warning sub-region, the equipment density is reduced according to the decreasing priority, the inner layer peripheral sub-region equipment density is 1 / 2 of the core area, and the outer layer peripheral sub-region is 1 / 3 of the core area; S22: The wireless radio spectrum monitoring module uses frequency hopping scanning technology to capture real-time wireless radio spectrum monitoring data including unmanned aerial vehicle communication frequency band signals; The millimeter wave radar array realizes dynamic tracking of low, slow and small targets through adaptive beam forming algorithm, and collects millimeter wave radar data including real-time three-dimensional coordinates (longitude X, latitude Y, height H) of the unmanned aerial vehicle; The light and electrical tracking equipment obtains light and electrical data containing visual positioning information of the unmanned aerial vehicle, including azimuth angle and distance of the unmanned aerial vehicle relative to the reference object, and assists in verifying the accuracy of the radar coordinates. When the radar signal is blocked, the light and electrical coordinates are automatically switched as the matching basis for the sub-region; The acoustic sensor identifies the propeller noise of the unmanned aerial vehicle based on the voiceprint feature library to obtain acoustic sensor data. Each detection equipment realizes synchronous data transmission through 5G private network, and the data transmission bandwidth and processing priority of high priority sub-region are higher than those of low priority sub-region.

[0027] Embodiment 2 On the basis of embodiment 1, the specific process of step S3 is as follows: S31: Through coordinate calibration with the sub-region boundary coordinates preset in the defense region GIS geographic information system, the boundary includes the polygon / ring boundary longitude and latitude range of the core defense sub-region and the peripheral early warning sub-region, and the terrain elevation difference is eliminated. The height value can be corrected by digital elevation model DEM.

[0028] S32: Match the calibrated real-time coordinates of the UAV with the boundaries of each sub-region: If the coordinates fall within the boundary of the core defense sub-region, such as within the latitude and longitude range N30°00'-N30°05', E120°00'-E120°05', it is determined to be a core zone target; If the coordinates fall within the inner peripheral sub-region, i.e. within the 1-3 km annular range outside the core zone, it is determined to be an inner peripheral target; If the coordinates fall within the outer peripheral sub-region, i.e. within the 3-10 km annular range outside the core zone, it is determined to be an outer peripheral target; Add a sub-region attribution label to each target UAV, and call the corresponding sub-region's no-fly parameters; S33: Real-time comparison of UAV data with sub-region attribution labels from three dimensions: height, speed, and stay duration: Height comparison: Compare the height value H collected in real-time by the millimeter wave radar with the sub-region height threshold, such as core region H>0m, inner peripheral H>50m, i.e. determine that the height is over-standard; Speed comparison: Calculate the instantaneous speed V of the UAV in km / h through the Doppler effect of the radar, and compare it with the sub-region speed threshold, such as inner peripheral V>30km / h, i.e. determine that the speed is over-standard; Stay duration comparison: Record the time T1 when the UAV first enters the sub-region, and calculate the stay duration ΔT=T 当前 - T1, and compare it with the sub-region stay duration threshold, such as outer peripheral ΔT>15min, i.e. determine that the stay duration is over-standard; If any of the over-standard conditions in the three dimensions is met, the core zone target is determined to be over-standard as soon as it enters, i.e. marked as an over-limit target, and the original data including coordinates, speed, height, and timestamp are included in the over-limit target dataset; data that is not over-standard is marked as compliant target and only the basic record is kept, without entering the subsequent feature extraction link.

[0029] S34: Preprocess the over-limit target dataset to eliminate interference data: Radio spectrum monitoring module data: Remove clutter signals within the frequency band, such as civilian broadcast signals and other electronic device interference signals, and only keep UAV communication signals with power ≥-110dBm through signal power threshold screening. Acoustic sensor data: Use wavelet transform denoising algorithm to filter environmental noise such as wind noise and vehicle noise, and retain the characteristic frequency band of 200-2000Hz of UAV propeller noise; Radar and photoelectric data: Remove outliers such as instantaneous height jumps and speed outliers caused by signal reflection, and use a sliding average method with a window size of 5 sampling points to smooth the data to ensure data continuity.

[0030] Temporal and spatial alignment of data: Due to the difference in sampling frequency of different detection devices, such as radar sampling interval 10ms, photoelectric sampling interval 30ms, through timestamp synchronization algorithm, taking the unified clock of 5G private network as the reference, the multi-source data is aligned to the same time dimension, such as generating 1 fusion data record every 10ms. At the same time, the spatial coordinates of each device are unified into the local coordinate system of the defense area, such as the plane rectangular coordinate system with the center point of the core area as the origin, to ensure the spatial consistency of the data.

[0031] S35: Based on the preprocessed hyper-limit target data set, the features of each type of data are extracted respectively to generate the target feature set of each sub-region.

[0032] The feature extraction result of each hyper-limit target generates a structured feature set according to the type of sub-region, and the format is as follows: Core defense sub-region feature set: {target number: 001, sub-region label: core area, speed feature: [V 均值 =25km / h, ΔV / Δt=3km / h・s], height feature: [H max =10m, over-limit duration=12s], communication protocol: encrypted private protocol, RCS: 0.01㎡, voiceprint abnormality: 85 points (full score 100)}; The feature set is accompanied by sub-region identification, such as core area feature set-20250925-001, which is transmitted in real time to the target dynamic classification model as the core input data for threat level judgment; at the same time, the feature set of each sub-region is stored in the database in time sequence for subsequent management effect evaluation and intrusion rule analysis.

[0033] In this embodiment, the process of extracting features of each type of data in step S35 is as follows: For millimeter wave radar data, instantaneous speed is calculated, flight speed features including mean and variance of speed are extracted, and RCS value is calculated based on radar echo power, unit:㎡, radar reflection cross section RCS features including mean and fluctuation range of RCS are extracted.

[0034] For millimeter wave radar data and photoelectric data, radar height H1 and photoelectric height H2 are obtained, the weight ratio of which is 6:4, flight height features including maximum height and continuous over-limit duration are extracted.

[0035] For radio spectrum monitoring data, communication protocol type features are extracted by analyzing signal modulation method, frame structure, and matching protocol library including WiFi, Bluetooth, ZigBee, and private protocol.

[0036] For acoustic sensor data, the frequency spectrum peak value, bandwidth, and period of noise signal are extracted, and the voiceprint information features are obtained by matching with the drone voiceprint library.

[0037] Embodiment 3 Based on Embodiment 1 or Embodiment 2, referring to Figure 3 , the target dynamic classification model is based on a traditional CNN, with the addition of a sub-region feature attention module and a multi-dimensional feature fusion layer, and the architecture is divided into 3 layers: Convolution layer (Conv Layer): Three convolution kernels with sizes of 3x1, 5x1, and 7x13 are used to extract features from the standardized input vector, capture local feature correlations, including the cooperative relationship between the speed change rate and the high approximation rate, and output a 64-dimensional feature map.

[0038] Attention layer (Attention Layer): Based on the dynamic allocation of feature weights by sub-region ownership labels, the core area feature set is given twice the weight of the communication protocol encryption level and the voiceprint abnormality, with a weight coefficient of 2.0, and the peripheral area is given 1.5 times the weight of the speed exceeding amplitude and the stay exceeding time length, to ensure that high-priority area key features are identified first.

[0039] Fully connected layer (FC Layer): The feature maps output by the convolution layer and the attention layer are flattened into 128-dimensional vectors, and non-linear transformation is performed through two fully connected layers with 64 and 32 hidden units, respectively, to finally output a 3-dimensional vector corresponding to the probability values of the three threat levels.

[0040] In this embodiment, the specific process of step S4 is as follows: S41: Structurally convert the target feature set into a numerical matrix format that can be recognized by the target dynamic classification model: For example, a certain core area feature set in the target feature set output by step S3 is: Core area feature set: {Target number: 001, speed feature: [V 均值 = 25 km / h, ΔV / Δt = 3 km / h・s], height feature: [H max = 10 m, exceeding time length = 12 s], communication protocol: encrypted private protocol, RCS: 0.01 m2, voiceprint abnormality: 85 points} is converted into a numerical matrix format that can be recognized by the model: Direct quantification of numerical features: The continuous numerical values of speed average (25), speed change rate (3), height maximum (10), exceeding time length (12), RCS value (0.01), and voiceprint abnormality (85) are directly retained as original data; Encoding conversion of type features: The communication protocol type is converted into a binary vector using one-hot encoding, such as encrypted private protocol → [1, 0, 0, 0], WiFi → [0, 1, 0, 0], to ensure that the model can perform numerical operations. Generate input vector: arrange all features in a fixed order of speed feature → height feature → communication protocol feature → RCS feature → voiceprint feature, forming an input vector with a dimension of 1 × N, N is the total number of features, such as the core area feature set N = 8, the peripheral area N = 6.

[0041] S42: To eliminate the influence of different feature dimension differences on model inference, such as speed unit km / h and RCS unit m2, Z-Score standardization algorithm is used to normalize the input vector: Calculation formula: x norm =( x - μ ) / σ , where x is the original feature value, μ is the mean of the feature in the training data set, σ is the standard deviation.

[0042] Standardization range: map all feature values to the interval [-1, 1], for example, the original value of voiceprint anomaly degree is 85, μ = 50, σ = 20, after standardization, it is (85−50) / 20 = 1.75, and then it is limited to 1 through truncation processing to ensure the stability of model input data.

[0043] Sub-area differentiation parameters: calculate the standardization parameters for the core area and peripheral area feature sets respectively, such as the core area RCS training set μ = 0.008 m2, and the peripheral area μ = 0.015 m2, to avoid errors caused by differences in cross-area feature distribution.

[0044] S43: Local feature extraction (convolution layer): taking a target feature set in the core area as an example, the convolution kernel calculates the local feature response through a sliding window with a step of 1, for example, a 3 × 1 convolution kernel calculates the associated features of "speed average - speed change rate - height maximum value", and the output feature value reflects the degree of cooperation of the three, such as a sudden increase in speed and a close approach to the core area, the feature value will significantly increase. Attention weight adjustment: the model generates an attention weight matrix through the Sigmoid function, and the core area target "communication protocol encryption level" feature weight calculation is w= σ (a × x + b), a and b are feature weight coefficients, a = 0.8, b = 0.2, x is the sub-area identification vector, if it is a core area target, x = 1, then w= σ (0.8 × 1 + 0.2) = 0.73, multiplied by the basic weight 2.0, the final weight is 1.46.

[0045] Gradient backpropagation optimization: During the model training phase, the error between the predicted value and the true label (manually labeled drone threat level) is calculated using the cross-entropy loss function. The Adam optimizer (learning rate 0.001) is used to backpropagate and adjust the parameters of each layer to ensure that the prediction accuracy of the core area threat level is ≥99.5% and the outer area is ≥99.2%.

[0046] S44: Generating threat level probability values ​​using the Softmax function: The 3D vector output by the fully connected layer (e.g., [0.12, 0.35, 0.53]) is converted into probability values ​​for each threat level using the Softmax function, satisfying a probability sum of 1: Calculation formula: ,in zi The first output of the fully connected layer i A vector value, where j is the category number of the threat level, j=1, 2, 3, zj The original fraction; For example: when the output of the fully connected layer is [0.2, 0.5, 0.8], the probabilities of each level are: Harmless civilian type P1= e 0.2 / ( e 0.2 + e 0.5 + e 0.8 ) ≈1.22 / (1.22+1.65+2.23)≈23%; Suspicious reconnaissance type P2 = 1.65 / 5.1 ≈ 32%; High-risk attack type P3 = 2.23 / 5.1 ≈ 44%.

[0047] S45: The threat level is determined by combining the maximum probability priority rule with the sub-regional differentiation threshold rule (see Table 1), and the final threat level is output: Table 1 Set up special case handling strategies: If a target in the core area uses a dedicated encrypted protocol and has an abnormal voiceprint score of ≥80, or ≥1.0 after standardization, it will be forcibly upgraded to a suspicious reconnaissance target even if P3 <0.4. If a target in the outer area exceeds the speed limit by ≥50%, such as an inner-outer speed of ≥45km / h, the P2 weight will be increased by 0.1 to avoid misjudgment.

[0048] Setting up the optimization process for the target dynamic classification model: New drone feature adaptive learning: When the model encounters an extracted unknown feature set, such as the unique communication protocol of a new drone or RCS features, initiate a real-time learning process: Feature anomaly detection: By calculating the Euclidean distance between the input features and the training set features, such as the RCS value deviating more than 3σ from the training set standard deviation, it is determined that the feature is unknown, and the target is marked as a learning sample.

[0049] Edge computing fine-tuning: Generate a small number of labeled samples on local edge computing nodes, combined with manual annotation, such as new suspicious drones, use transfer learning technology to freeze the model convolution layer parameters, and only fine-tune the fully connected layer and attention layer parameters.

[0050] Update cycle: Trigger update when the number of core area learning samples accumulates ≥5, cycle ≤30 minutes, peripheral area ≥10, cycle ≤1 hour, and after updating, the model is automatically synchronized to each sub-area detection node.

[0051] Historical data feedback optimization: Daily early morning analysis of the feature set-threat level-actual control result data of the previous 24 hours: Error statistics: Calculate the misjudgment rate of each sub-area threat level, such as the proportion of high-risk attack type misjudged as suspicious reconnaissance type in the core area, if the misjudgment rate ≥0.5%, adjust the attention weight of the corresponding sub-area, such as increasing the RCS feature weight in the core area; Threshold correction: If a certain type of drone (such as small folding reconnaissance drones) is frequently misjudged in the core area, recalculate the Softmax decision threshold of the target feature of this type, such as reducing the P3 decision threshold from 0.4 to 0.35, to ensure that the model adapts to the actual application scenario.

[0052] The specific process of step S5 is as follows: S51: Set the correspondence between sub-area priority and resource configuration, see Table 2 below: Table 2 Set the priority-threat level control strategy decision matrix, see Table 3 below: Table 3 S52: Perform differentiated operations according to priority-threat level matching: Control of harmless civilian drones (low-intensity control): For civilian drones without carrying threat devices and only slightly exceeding the standard, implement control measures mainly for early warning and guidance, the specific process is as follows: The execution process in the core area is as follows: After receiving the harmless civilian type result of the model output, the geo-fencing system is started within 1s to generate the core area boundary electronic fence.

[0053] Send the drone a Chinese / English bilingual instruction containing the coordinates of the no-fly zone and the consequences of violating the instruction through the ADS-B system (Automatic Dependent Surveillance-Broadcast). The instruction is sent at a frequency of 1 time per second to ensure that the operator receives it.

[0054] The photoelectric tracking device continuously locks the drone and monitors whether it leaves according to the instruction in real time. If it does not leave within 30 seconds, it is automatically upgraded to a suspicious reconnaissance type control. The outer peripheral area executes the following process: The inner peripheral area: within 2 seconds, send a warning instruction through ADS-B, and track the noise of the drone using acoustic sensors (sample rate increased to 48kHz) to determine its flight direction. The outer peripheral area: within 3 seconds, send a warning SMS to the operator's mobile phone through the drone control platform, and record the drone's location (updated every 5 seconds). If it moves towards the core area, trigger the inner area warning. Suspicious reconnaissance type drone control (medium intensity control): For drones with reconnaissance features (such as carrying high-definition cameras, encrypted communication protocols), perform a combination of interference + locking + recording control. The specific steps are as follows: Core area execution process: within 1 second, activate the high-power electromagnetic interference module, generate a directional interference signal based on the drone's communication frequency band provided by the radio spectrum monitoring module, with a power of 8-10W. The interference direction is calibrated in real time by a millimeter wave radar with a deviation of ≤1°. The millimeter wave radar and photoelectric device are dual-mode locked, recording the drone's flight trajectory, including latitude, longitude, altitude, and speed, storing 1 data every 100ms. Simultaneously notify the ground control personnel (equipped with portable interference devices) to go to the possible landing area of the drone. If the drone still stays after interference, within 5 seconds, start the anti-drone drone group on standby; Outer peripheral area execution process: Inner peripheral area: within 2 seconds, start medium-power interference (3-5W), focusing on interfering with the drone's image transmission frequency band (such as 5.8GHz), while analyzing whether the flight trajectory has evasive characteristics (such as frequent changes in direction). Outer peripheral area: within 3 seconds, start low-power interference (0.1-2W), only interfere with the remote control frequency band (without affecting surrounding civilian equipment), and upload the trajectory data to the regional defense platform to predict whether it is moving towards the core area. High-risk attack type drone control (high intensity control): For drones carrying attack devices (such as explosives, laser emitters) or with high-speed penetration characteristics, perform interception + linkage + emergency all-around control. The specific steps are as follows: Core area execution process: within 1s, trigger anti-drone drone swarm take-off command, 3-5 intercept drones take off from the preset landing point (5 around the core area, spacing ≤1km), calibrate the target position through laser radar (detection distance ≥1km). Intercept drones use triangular encirclement tactics to form an intercept circle with a diameter of 50m, when the distance to the target is ≤50m, release the foldable capture net, the capture success rate ≥99%. Synchronous linkage of the first-level ground air defense system (such as short-range air defense missiles) enters the standby state, if the capture fails, start the air defense missile locking program within 10s, and evacuate the personnel in the core area; Peripheral area execution process: inner peripheral area: within 2s, dispatch 1-3 intercept drones, use tail-chasing interception tactics, electromagnetic interference module (3-5W) continuously interferes with the target navigation signal (GPS / Beidou), forcing the target to slow down. Outer peripheral area: within 3s, call the standby intercept drone, and send a request for assistance to the adjacent inner area, share target data through blockchain technology, form a cross-regional interception echelon, if the target breaks through the inner defense line, automatically trigger the core area interception plan. Cross-regional cooperative linkage stage: in response to multiple targets, cross-sub-regional intrusion scenarios: Multi-target simultaneous intrusion coordination mechanism: When multiple sub-regions simultaneously detect high-risk attack drones, such as 1 in the core area, 2 in the inner peripheral area, start the priority priority, resource allocation coordination process: The core area occupies 3 intercept drones, the inner peripheral area calls the remaining 2, and at the same time, 1 standby drone from the outer peripheral area is allocated to support the inner layer. The electromagnetic interference module is uniformly scheduled by the regional control center to avoid the same frequency interference conflict, such as core area interference 2.4GHz, inner layer interference 5.8GHz. Use time slicing technology to allocate radar resources, core area radar scans every 5ms, inner layer scans every 10ms, to ensure the continuity of high-priority area monitoring. Cross-sub-regional moving target tracking mechanism: when the drone moves from the outer peripheral area to the core area (crosses the sub-regional intrusion), execute the relay control process: The outer peripheral area synchronizes the target's flight trajectory and characteristic parameters (RCS, voiceprint) to the inner peripheral area through the 5G private network when the target leaves, the synchronization time ≤1s; The inner peripheral area preheats the electromagnetic interference module (power adjusted to 5W) and starts the standby of the intercept drone, when the target enters the inner area, directly executes the medium-level control, without the need to re-identify; If the target breaks through the inner area, the core area automatically upgrades the control strength (such as increasing the number of intercept drones to 5), forming a three-level defense line of outer warning-inner interception-core defense.

[0055] The specific process of step S6 is as follows: S61: Set differentiated evaluation indexes for different control measures to ensure that the control effect can be quantified: Early warning index: harmless civilian drone departure rate, core area ≥98%, inner layer ≥95%, outer layer ≥90%, and instruction reception success rate ≥99%; Interference index: suspicious reconnaissance drone image transmission interruption rate, core area ≥99%, inner layer ≥97%, outer layer ≥95%, and false interference rate ≤0.1%; Interception index: high-risk attack drone capture success rate, core area ≥99%, inner layer ≥98%, outer layer ≥97%, and interception response time ≤1s / 2s / 3s.

[0056] S62: According to the evaluation results, adjust the control strategy and resource allocation in reverse, and the specific optimization direction is: If the core area high-risk attack drone interception success rate is lower than 99%, increase the number of interception drones from 3-5 to 5-7, or increase the capture net deployment diameter from 4m to 5m; If the inner layer outer area suspicious reconnaissance drone false interference rate is higher than 0.1%, optimize the electromagnetic interference waveform generation algorithm, and increase the identification and filtering of civilian equipment signals, such as excluding mobile phone communication frequency bands.

[0057] If the outer layer outer area harmless civilian drone departure rate is lower than 90%, upgrade the early warning method from SMS warning to APP pop-up window + phone reminder, and expand the early warning range from 1km outside the outer layer outer area boundary to 2km.

[0058] S63: Set up an emergency plan update mechanism: Based on historical control data, such as new drone intrusion cases and control failure events, update the priority-threat level decision matrix every month: If a new type of drone (such as a micro-bionic drone) is found to be easily misjudged as harmless in the core area, increase its RCS feature weight from 1.0 to 1.5, and lower the high-risk attack type judgment threshold from P3≥0.4 to P3≥0.35.

[0059] If multiple target cooperative intrusion occurs in the outer peripheral area, add a new peripheral area multi-target joint plan, preset the adjacent area resource allocation path, and shorten the support response time from 3s to 2s.

[0060] The unmanned aerial vehicle detection and control system for regional defense is used for realizing the unmanned aerial vehicle detection and control method for regional defense, and comprises a region division and parameter setting module, a multi-dimensional cooperative detection network module, a target data processing and feature extraction module, a target dynamic classification model, a hierarchical dynamic control execution module, and a control effect evaluation and strategy optimization module. Each module realizes data interaction and cooperative work through a 5G private network or an optical fiber network.

[0061] The region division and parameter setting module is used for spatially dividing a region to be defended and presetting parameters. The multi-dimensional cooperative detection network module is based on the output of the region division and parameter setting module, differentiates detection devices in each sub-region, and is used for collecting multi-source data of a target unmanned aerial vehicle in real time. The target data processing and feature extraction module is used for receiving multi-source data transmitted by the multi-dimensional cooperative detection network module, performing data screening, preprocessing, and feature extraction. The target dynamic classification model is used for receiving target features from the target data processing and feature extraction module, and realizing unmanned aerial vehicle threat level determination. The hierarchical dynamic control execution module is used for executing differential control operations according to the response priority of the region division and parameter setting module and the threat level of the target dynamic classification model. The control effect evaluation and strategy optimization module is used for continuously collecting post-control data from the hierarchical dynamic control execution module, realizing effect evaluation and strategy adjustment.

Claims

1. A method for detecting and controlling unmanned aerial vehicles (UAVs) for regional defense, characterized in that, Includes the following steps: S1: Divide the area to be defended into at least one core defense sub-area and at least two peripheral early warning sub-areas, and set the drone no-fly threshold and detection response priority for each sub-area; S2: Construct a multi-dimensional collaborative detection network based on sub-region division: Detection equipment is deployed differently in each sub-region, and the detection equipment in each sub-region collects multi-source data of each target UAV in real time within the corresponding spatial range; S3: Based on multi-source data, determine the sub-region where the drone is located and whether it exceeds the no-fly threshold of that sub-region, obtain target data that exceeds the threshold, extract features from the target data, and obtain target feature sets for each sub-region, including flight speed, altitude, communication protocol type, radar cross section and acoustic information; S4: Input the extracted feature set into the target dynamic classification model, and the target dynamic classification model outputs the threat level of each target UAV in real time based on the feature set; S5: Implement a hierarchical dynamic control strategy based on the detection response priority of each sub-region and the corresponding target drone threat level; S6: After implementing the hierarchical dynamic control strategy, continuously collect the status parameters of each target UAV, dynamically evaluate the control effect based on the detection response priority of each sub-region and the preset evaluation indicators, and adaptively adjust the detection equipment parameters and control strategy of each sub-region according to the evaluation results.

2. The method for detecting and controlling unmanned aerial vehicles (UAVs) for regional defense according to claim 1, characterized in that, The specific process of step S1 is as follows: S11: Divide the area to be defended into spatial sub-areas, and determine at least one core defense sub-area and at least two peripheral early warning sub-areas; S12: Set no-fly thresholds for drones for each sub-region; S13: Set the detection response priority for each sub-region. The core defense sub-region has the highest priority, the inner outer sub-region has the medium priority, and the outer outer sub-region has the basic priority.

3. The method for detecting and controlling unmanned aerial vehicles (UAVs) for regional defense according to claim 2, characterized in that, The specific process of step S2 is as follows: S21: Based on each sub-region and its corresponding detection priority, detection equipment is deployed in each sub-region in a differentiated manner to build a multi-dimensional collaborative detection network, including radio spectrum monitoring modules, millimeter-wave radar arrays, photoelectric tracking equipment and acoustic sensors; S22: The radio spectrum monitoring module uses frequency hopping scanning technology to capture radio spectrum monitoring data, including UAV communication frequency band signals, in real time; Millimeter-wave radar arrays achieve dynamic tracking of low, slow, and small targets and acquire millimeter-wave radar data through adaptive beamforming algorithms; The photoelectric tracking device acquires photoelectric data containing the visual positioning information of the UAV; The acoustic sensor uses an acoustic signature database to identify the propeller noise of the drone and obtain acoustic sensor data.

4. The method for detecting and controlling unmanned aerial vehicles (UAVs) for regional defense according to claim 3, characterized in that, The specific process of step S3 is as follows: S31: Perform coordinate calibration by comparing the coordinates of the sub-regional boundaries with the preset coordinates of the GIS geographic information system of the defense area; S32: Match the calibrated UAV real-time coordinates with the boundaries of each sub-region: S33: For drone data with added sub-region affiliation tags, real-time comparison is performed from three dimensions: altitude, speed, and dwell time. If any of the dimensions exceeds the standard, it is marked as an over-limit target and its original data is included in the over-limit target dataset; data that does not exceed the standard is marked as a compliant target, only the basic record is retained, and it is not included in the subsequent feature extraction process. S34: Preprocess the dataset of out-of-limit targets, remove interfering data, and perform spatiotemporal alignment of the data; S35: Based on the preprocessed target dataset, extract the corresponding features from each type of data to generate the target feature set for each sub-region.

5. The method for detecting and controlling unmanned aerial vehicles (UAVs) for regional defense according to claim 4, characterized in that, The feature extraction process for each type of data in step S35 is as follows: Instantaneous velocity is calculated from millimeter-wave radar data, and flight velocity features including mean and variance are extracted. RCS values ​​are calculated based on radar echo power, and radar cross section RCS features including mean and fluctuation range are extracted. For millimeter-wave radar data and optoelectronic data, obtain radar altitude H1 and optoelectronic altitude H2, and extract flight altitude features such as maximum altitude and duration of continuous exceedance. For radio spectrum monitoring data, communication protocol type features are extracted by analyzing signal modulation methods and frame structures and matching protocol libraries. For acoustic sensor data, the peak frequency, bandwidth, and period of the noise signal are extracted and matched with the UAV voiceprint database to obtain voiceprint information features.

6. The method for detecting and controlling unmanned aerial vehicles (UAVs) for regional defense according to claim 1, characterized in that, The target dynamic classification model is based on traditional CNN, with the addition of a sub-region feature attention module and a multi-dimensional feature fusion layer. The architecture consists of three layers: Convolutional layer: Three convolutional kernels are used to extract features from the standardized input vector, capture local feature associations, and output a feature map; Attention layer: Based on the sub-region affiliation label, feature weights are dynamically assigned. The core area feature set is given 2 times the weight for communication protocol encryption level and voiceprint anomaly, while the outer area is given 1.5 times the weight for speed exceeding the standard and dwell time exceeding the standard, ensuring that key features in high-priority areas are identified first. Fully connected layer: The feature maps output by the convolutional layer and attention layer are flattened into a 128-dimensional vector, and then non-linearly transformed through two fully connected layers to finally output a 3-dimensional vector, corresponding to the probability values ​​of the three threat levels.

7. The method for detecting and controlling unmanned aerial vehicles (UAVs) for regional defense according to claim 1, characterized in that, The specific process of step S4 is as follows: S41: Perform a structured transformation on the target feature set, converting it into a numerical matrix format recognizable by the target dynamic classification model: S42: The input vector is normalized using the Z-Score normalization algorithm; S43: Taking a target feature set in the core area as an example, the convolutional kernel calculates the local feature response through a sliding window, and the model generates the attention weight matrix through the Sigmoid function; During the model training phase, the error between the predicted value and the true label is calculated using the cross-entropy loss function, and the Adam optimizer is used to adjust the parameters of each layer in reverse. S44: Threat Level Probability Calculation: The 3D vector output by the fully connected layer is converted into probability values ​​for each threat level using the Softmax function, satisfying a probability sum of 1: S45: Combines the maximum probability priority and sub-regional differentiation threshold dual rules to determine the threat level and outputs the final threat level.

8. The method for detecting and controlling unmanned aerial vehicles (UAVs) for regional defense according to claim 1, characterized in that, The specific process of step S5 is as follows: S51: Set the correspondence between sub-region priorities and resource allocation, and set the priority-threat level management strategy decision matrix. S52: Perform differentiated actions based on priority-threat level matching: Low-intensity control measures will be implemented for harmless civilian drones, medium-intensity control measures will be implemented for suspicious reconnaissance drones, and high-intensity control measures will be implemented for high-risk attack drones.

9. The method for detecting and controlling unmanned aerial vehicles (UAVs) for regional defense according to claim 1, characterized in that, The specific process of step S6 is as follows: S61: Set differentiated evaluation indicators for different control measures: Early warning indicators: Departure rate of harmless civilian drones, command reception success rate; Interference-related indicators: Image transmission interruption rate and false interference rate of suspicious reconnaissance UAVs; Interception metrics: Success rate of capturing high-risk attack drones, interception response time; S62: Based on the assessment results, adjust the control strategy and resource allocation in reverse: If the success rate of intercepting high-risk attack drones in the core area is less than 99%, increase the number of drones to be intercepted or increase the diameter of the capture net. If the false interference rate of suspicious reconnaissance drones in the inner and outer perimeter areas is higher than 0.1%, optimize the electromagnetic interference waveform generation algorithm and increase the identification and filtering of signals from civilian equipment; If the departure rate of harmless civilian drones in the outer perimeter area is less than 90%, the early warning method will be upgraded and the early warning range will be expanded. S63: Establish an emergency response plan update mechanism: The priority-threat level decision matrix is ​​updated monthly based on historical control data. If a new type of drone is found to be easily misjudged as harmless in the core area, its RCS feature weight will be increased, while the threshold for judging high-risk attack drones will be lowered. If multiple coordinated intrusions occur repeatedly in the outer perimeter area, a new multi-target linkage contingency plan for the outer perimeter area will be added, and a resource allocation path for adjacent areas will be preset to shorten the support response time.

10. A drone detection and control system for regional defense, used to implement the drone detection and control method for regional defense as described in any one of claims 1-9, characterized in that, It includes a region division and parameter setting module, a multi-dimensional collaborative detection network module, a target data processing and feature extraction module, a target dynamic classification model, a hierarchical dynamic control execution module, and a control effect evaluation and strategy optimization module; each module achieves data interaction and collaborative work through a 5G private network or fiber optic network; The area division and parameter setting module is used to spatially divide the area to be defended and preset the parameters. The multi-dimensional collaborative detection network module, based on the output of the region division and parameter setting module, deploys detection devices in different sub-regions to collect multi-source data of the target UAV in real time. The target data processing and feature extraction module is used to receive multi-source data transmitted by the multi-dimensional collaborative detection network module and perform data filtering, preprocessing and feature extraction. The target dynamic classification model is used to receive the target feature set from the target data processing and feature extraction module to determine the threat level of the UAV. The hierarchical dynamic control execution module is used to perform differentiated control operations based on the response priority of the regional division and parameter setting module and the threat level of the target dynamic classification module. The control effect evaluation and strategy optimization module is used to continuously collect post-control data from the hierarchical dynamic control execution module to achieve effect evaluation and strategy adjustment.

Images