A system and method for repelling target navigation by a specific frequency electromagnetic wave interference

Abstract

The application discloses a kind of specific frequency electromagnetic wave interference target navigation's expelling system and method, belong to navigation countermeasure technical field.The present application obtains nest coordinate set, obtains the stress threshold of the bird class to which each nest belongs;Calculate the noise influence range from inside to outside, form nest area influence range set;Record the model of target unmanned aerial vehicle and target coordinate;The target coordinate at the present time is matched with nest area influence range set in space, and the output judgment result is output;When target is in nest area influence range, output frequency is normal interference frequency;When target is in nest area influence range, whether the bird is in nest, whether environmental noise value is greater than stress threshold is judged, and sub-scene judgment and interference execution are carried out;In the process of navigation interference, screening is separated from boundary point, and predetermined trajectory is generated;Based on predetermined trajectory, target is interfered with navigation, and when the current coordinate of target is separated from bird breeding area, navigation interference is terminated.

Description

Technical Field

[0001] This invention relates to the field of navigation countermeasures technology, specifically a system and method for driving away targets whose navigation is interfered with by electromagnetic waves of a specific frequency. Background Technology

[0002] With the popularization of civilian drone technology, its application in fields such as photography and agriculture is becoming increasingly widespread. However, it also poses a serious threat to ecologically sensitive areas, especially during bird breeding season. Current drone countermeasures mainly involve cutting off drone navigation or communication links through radio frequency interference and GPS spoofing. While mainstream equipment can cover commonly used frequency bands and achieve rapid suppression, it has significant limitations in the special scenario of bird breeding grounds, making it difficult to balance the effectiveness of deterring drones with the needs of ecological protection.

[0003] Traditional countermeasures devices employ broad-spectrum or fixed-frequency jamming modes, failing to consider the physiological differences in birds during their breeding season. Different birds exhibit significantly different stress thresholds for noise, and the tolerance of the same bird varies depending on whether it is nesting or not. Current technologies can only achieve blinding interference, not guide drones to directional escape. Jammed drones may fall randomly, potentially causing secondary interference or even physical damage if they crash into the influence range of other nests, chick activity areas, or ecologically sensitive zones. Furthermore, existing technologies often treat the entire breeding ground as a single control unit, without differentiating nesting areas. The distribution and sensitivity range of different nests within a bird's breeding ground vary, meaning existing devices may over-interfere with targets far from nests or under-interfere with targets close to core nesting areas. Summary of the Invention

[0004] The purpose of this invention is to provide a system and method for driving away targets whose navigation is interfered with by electromagnetic waves of a specific frequency, so as to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] In a first aspect, this application provides a method for driving away navigation targets that are interfered with by electromagnetic waves of a specific frequency, comprising the following steps:

[0007] The three-dimensional coordinates of all bird nests within the bird breeding grounds are obtained to form a set of bird nest coordinates; the bird species to which each nest belongs is identified through AI image recognition algorithms, and its stress threshold for noise during the breeding season is obtained; with each bird nest coordinate as the center point, the noise impact range from the inside out is calculated based on the stress threshold, forming a set of nest area impact ranges.

[0008] The system scans targets entering bird breeding areas, analyzes their features using visual recognition algorithms, matches them against a drone model database, records the drone's model, obtains the target's coordinates, and updates its trajectory.

[0009] Spatial matching is performed between the current target coordinates and the set of nest area influence ranges, and the judgment result is output. When the target is outside the nest area influence range, electromagnetic waves are generated, and the output frequency is the normal interference frequency to perform navigation interference. When the target is within the nest area influence range, the environmental noise value of the current associated nest area is obtained, and infrared thermal imaging is used to determine whether there are birds in the nests corresponding to the current associated nest area. The environmental noise value is judged to be greater than the stress threshold and whether there are birds in the nests, and sub-scene judgment and interference execution are performed.

[0010] During the navigation interference process, points that deviate from the boundary are selected and a predetermined trajectory is generated; the target trajectory is tracked, and navigation interference is carried out on the target based on the predetermined trajectory. When the target's current coordinates deviate from the bird breeding area, the navigation interference terminates.

[0011] In conjunction with the first aspect, in the first embodiment of the first aspect of this application, the step of identifying the bird species to which each nest belongs through an AI image recognition algorithm and obtaining its stress threshold for noise during the breeding season includes:

[0012] This study collects nest morphology features and surrounding activity characteristics of potential bird species in bird breeding grounds. Nest morphology features include nest size, shape, material, and nesting location characteristics. Surrounding activity characteristics include bird size, feather color, and behavioral patterns. Each nest morphology image and each video frame of bird activity is labeled with a corresponding bird category label, and key feature regions are marked to form a labeled dataset. For nest morphology feature recognition, a convolutional neural network (CNN) architecture is used. Convolutional layers extract spatial features of the nest, and fully connected layers output morphological feature vectors. For surrounding bird activity feature recognition, a hybrid architecture combining CNN and LSTM is used. First, the CNN extracts the appearance features of a single-frame bird image, then the LSTM processes multiple frames of temporal data to capture the dynamic features of bird activity, outputting an activity feature vector. The morphological feature vector and the activity feature vector are fused according to weights and input into a Softmax classifier to output a bird category probability distribution.

[0013] The labeled dataset is divided into training, validation and test sets for model training and optimization. Nest morphology features and surrounding activity features of the nests to be identified are collected and input into the model to obtain the bird category of the nests to be identified and obtain the stress threshold of noise during the breeding season.

[0014] In conjunction with the first aspect, in the second embodiment of the first aspect of this application, the step of calculating the noise impact range from the inside out based on the stress threshold, using the coordinates of each bird's nest as the center point, to form a set of nest area impact ranges, includes:

[0015] Measure environmental parameters around the bird's nest, including the terrain attenuation coefficient k and the weather correction factor. vegetation density correction factor The point source spherical wave attenuation model is adopted, and the basic formula is: Where L(r) is the noise value at a distance of r meters from the interference source. To determine the noise source intensity when the interference system outputs electromagnetic waves at the normal interference frequency, adjustments are made according to the different drone models. r is the straight-line distance from the interference source to the target point, and 20lg(r) is the geometric attenuation term for the spherical wave. The normal interference frequency is obtained by calling a preset mapping table between drone models and normal interference frequencies, based on the drone model. The basic formula is then modified to obtain the actual noise propagation formula, specifically including: the downwind noise value in the downwind direction. Headwind noise level In vegetated areas, the noise level of vegetation cover is... When there are multiple types of terrain or vegetation around the bird's nest, the annular area centered on the bird's nest is divided according to the terrain type, and different areas are calculated using the corresponding k value.

[0016] Solve for r such that ,in, This represents the lower limit of the stress threshold, and r is calculated by back-calculating using the actual noise propagation formula; a noise meter is deployed at the calculated r point, and the interference device is turned on until... Measure the actual noise value; when When k is increased, r is recalculated, and the measurement is repeated until... The influence range of all bird nests is summarized to form a set of influence ranges of nest areas.

[0017] In conjunction with the first aspect, in the third embodiment of the first aspect of this application, the step of scanning targets entering bird breeding areas, analyzing target features using a visual recognition algorithm, matching them against a drone model database, and recording the drone model includes:

[0018] The radar scans the bird breeding area along a preset route, outputs a list of moving targets in real time, sets a threshold for UAV motion characteristics to exclude non-UAV targets, preprocesses the acquired images to obtain a set of feature images to be identified, calls a convolutional neural network algorithm to extract features from the set of feature images to be identified, including structural features, detailed features, and thermal features of the targets, calls a UAV model feature database, uses cosine similarity to calculate the similarity between the feature vector of the target to be identified and the standard feature vector of each model in the database, matches the UAV model and records it.

[0019] In conjunction with the first aspect, in the fourth embodiment of the first aspect of this application, the step of spatially matching the current target coordinates with the set of influence ranges of the nesting area and outputting the judgment result includes:

[0020] The Euclidean distance formula is used to calculate the planar distance between the target and the center point of the nest area. Based on the set of influence ranges of the nest area, it is determined whether the target coordinates fall within the nest area. When the current nest area overlaps with other nest areas, it is simultaneously checked whether the target is in another overlapping nest area. When the target is in the range of multiple nest areas at the same time, the target is marked as being in the overlapping area of ​​nest areas. According to the distribution of nest areas from the core area of ​​the breeding ground to the outer area, spatial matching is performed on each valid nest area in turn, and the judgment result is output.

[0021] In conjunction with the first aspect, in the fifth embodiment of the first aspect of this application, the step of generating electromagnetic waves with an output frequency of normal interference frequency when the target is outside the influence range of the nest area to perform navigation interference includes:

[0022] Based on the current distance between the target and the jamming device, combined with the normal jamming frequency, the output power of the power amplifier is set. According to the target coordinates and the position of the jamming device, the azimuth and elevation angles of the target relative to the jamming device are calculated, and the beam direction of the directional transmitting antenna is adjusted so that the main beam of the antenna is aimed at the target.

[0023] In conjunction with the first aspect, in the sixth embodiment of the first aspect of this application, the step of obtaining the environmental noise value of the currently associated nesting area when the target is within the influence range of the nesting area, and determining whether there are birds in the nests corresponding to the currently associated nesting area through infrared thermal imaging monitoring, includes:

[0024] The sensor is activated to continuously collect noise data, receiving and storing the environmental noise value every second in real time, and preprocessing the environmental noise value; the infrared thermal imager is activated to collect static thermal images of the bird's nest and surrounding area to determine whether there is a heat source area in the bird's nest area with a temperature higher than the ambient temperature; the dynamic video acquisition mode is switched to determine whether there are dynamic changes in the heat source area and to exclude non-bird heat sources; when there is a heat source area in the static thermal image that matches the characteristics of birds, and bird activity characteristics are observed in the area during dynamic monitoring, it is determined that there are birds in the bird's nest in the currently associated nest area.

[0025] In conjunction with the first aspect, in the seventh embodiment of the first aspect of this application, the step of determining whether the environmental noise value is greater than the stress threshold and whether the bird is in the nest, and performing sub-scene judgment and interference execution, includes:

[0026] In the first sub-scenario, the ambient noise level exceeds the stress threshold and birds are in their nests. The normal interference frequency corresponding to the target is obtained, and the minimum power to suppress the target's navigation system is set based on the distance between the target and the interfering device. In the second sub-scenario, the ambient noise level exceeds the stress threshold and birds are in their nests. The frequency is set the same as in the first sub-scenario, increasing the power based on the first sub-scenario setting, ensuring no excessive out-of-band radiation and no interference with nearby legitimate devices. In the third sub-scenario, the ambient noise level is not greater than the stress threshold and birds are in their nests. An initial interference frequency is set, lower than the normal interference frequency corresponding to the target, to ensure initial... If the noise generated by the interference signal, when superimposed with the current ambient noise, still does not exceed the stress threshold, a minimum safe power is set, lower than the power of the first scenario, to avoid the initial power being too high and causing the noise to exceed the stress threshold. When the target is not moving, the interference frequency is increased by a set ratio. If the superimposed noise still does not exceed the stress threshold, the new frequency is maintained. If it exceeds the threshold, it falls back to the previous frequency and is no longer increased. The increased interference frequency must not exceed the normal interference frequency corresponding to the target. In the fourth sub-scenario, the ambient noise value is not greater than the stress threshold and the birds are not in their nests. The normal interference frequency corresponding to the target is used, and the power is set to that of the first sub-scenario.

[0027] In conjunction with the first aspect, in the eighth embodiment of the first aspect of this application, the step of filtering out boundary points and generating a predetermined trajectory during the navigation interference process includes:

[0028] The selection criteria for points that have deviated from the boundary are ranked from highest to lowest as follows: points that are on the boundary of the breeding grounds; the straight-line distance from the point to the safe landing zone is relatively shorter among all boundary points; the line connecting the target's current location to the point passes through the fewest nesting areas; and there are no ecologically sensitive areas in or around the point.

[0029] The boundary line of the breeding ground is decomposed into several continuous boundary points, which are then selected according to screening criteria to obtain the boundary-free points. The target coordinates are used as the starting point of the predetermined trajectory, and the boundary-free points are used as the ending point of the predetermined trajectory. The starting point and the ending point of the trajectory are connected to form an initial straight-line trajectory. Trajectory constraints are set, including: the entire trajectory does not enter any nesting area; the deviation between the total trajectory length and the straight-line distance from the starting point to the ending point is minimized; the trajectory does not cross ecologically sensitive areas or the protection range of legal equipment.

[0030] When the initial straight trajectory fully conforms to the trajectory constraints, it is taken as the predetermined trajectory. When the initial straight trajectory crosses any region in the trajectory constraints, the path inflection point that bypasses the region with the least increase in distance is found on both sides of the crossed region to form the initial predetermined trajectory. The inflection point of the initial predetermined trajectory is smoothed. Based on the target's flight capability, it is determined whether the smoothed initial predetermined trajectory is within the target's executable range. If it is outside the range, the position of the trajectory inflection point is further adjusted until the trajectory conforms to the target's flight capability and the predetermined trajectory is formed.

[0031] Secondly, this application provides a system for driving away targets whose navigation is interfered with by electromagnetic waves of a specific frequency, comprising:

[0032] Nesting Area Impact Range Generation Module: Includes: Nest Coordinate Positioning Unit, Stress Threshold Acquisition Unit, and Nesting Area Impact Range Generation Unit; wherein, the Nest Coordinate Positioning Unit acquires the three-dimensional coordinates of all nests within the bird breeding grounds, forming a nest coordinate set; the Stress Threshold Acquisition Unit uses AI image recognition algorithms to identify the bird species to which each nest belongs and acquires its stress threshold for noise during the breeding season; the Nesting Area Impact Range Generation Unit uses each nest coordinate as the center point and, based on the stress threshold, calculates the noise impact range from the inside out, forming a nesting area impact range set;

[0033] Target analysis module: includes a drone model acquisition unit and a target coordinate acquisition unit; wherein, the drone model acquisition unit scans targets entering the bird breeding area, analyzes the target features using a visual recognition algorithm, matches them with the drone model database, and records the drone model; the target coordinate acquisition unit acquires the target coordinates and updates its movement trajectory;

[0034] The navigation interference execution module includes: a spatial matching unit, an out-of-range interference execution unit, an in-range interference data acquisition unit, and an in-range interference execution unit. The spatial matching unit performs spatial matching between the current target coordinates and the nest area influence range set, outputting the judgment result. The out-of-range interference execution unit generates electromagnetic waves when the target is outside the nest area influence range, outputting the normal interference frequency for navigation interference. The in-range interference data acquisition unit acquires the environmental noise value of the currently associated nest area when the target is within the nest area influence range, and uses infrared thermal imaging to determine whether there are birds in the corresponding nests. The in-range interference execution unit determines whether the environmental noise value exceeds the stress threshold and whether birds are in the nests, performing sub-scene judgment and interference execution.

[0035] The predetermined trajectory generation and execution module includes a predetermined trajectory generation unit and a predetermined trajectory execution unit. The predetermined trajectory generation unit filters out points that deviate from the boundary during the navigation interference process and generates a predetermined trajectory. The predetermined trajectory execution unit tracks the target trajectory and performs navigation interference on the target based on the predetermined trajectory. When the target's current coordinates deviate from the bird breeding area, the navigation interference terminates.

[0036] Compared with the prior art, the beneficial effects of the present invention are:

[0037] 1. This invention uses AI image recognition to identify the bird species in each nest to obtain a unique stress threshold, and also calculates the impact range of the nest area based on the stress threshold. When interfering within the nest area, it further combines the environmental noise value and the birds' nest status to perform interference in different scenarios, rather than using a uniform interference mode. This can accurately match the birds' tolerance and avoid interference signals exceeding the stress threshold and causing ecological risks.

[0038] 2. This invention filters out points that deviate from the boundary and generates a predetermined trajectory during the interference process. By tracking the target trajectory and implementing interference based on the predetermined trajectory, it ensures that the target moves out of the breeding ground along a preset safe path. The target is only allowed to fall after it leaves the breeding ground area, thus completely avoiding the problem of the target randomly falling into sensitive areas such as the nesting area and the chick activity area.

[0039] 3. This invention first constructs a set of nest area influence ranges, and then determines whether the target is inside or outside the nest area by matching the current target coordinates with the set. Then, it executes corresponding interference strategies to achieve differentiated control based on the nest area, taking into account both the effectiveness of interference and the accuracy of control. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the steps of a method for driving away navigation targets that are interfered with by electromagnetic waves of a specific frequency, according to the present invention.

[0041] Figure 2 This is a system structure diagram of a target navigation deterrent system that uses electromagnetic waves of a specific frequency to interfere with the target, according to the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] Example: Figures 1-2 As shown, the present invention provides a technical solution:

[0044] like Figure 1 As shown, this application provides a method for driving away targets whose navigation is interfered with by electromagnetic waves of a specific frequency, including the following steps:

[0045] Step S100: Obtain the three-dimensional coordinates of all bird nests within the bird breeding grounds to form a set of bird nest coordinates; use AI image recognition algorithms to identify the bird species to which each nest belongs and obtain its stress threshold for noise during the breeding season; take each bird nest coordinate as the center point and calculate the noise impact range from the inside out based on the stress threshold to form a set of nest area impact ranges.

[0046] Specifically, nest morphology features and surrounding activity features of potential bird species in bird breeding grounds are collected. Nest morphology features include the size, shape, material, and location of the nest. Surrounding activity features include the bird's body size, feather color, and behavioral patterns. Each nest morphology image and each video frame of bird activity is labeled with a corresponding bird category label, and key feature regions are marked to form a labeled dataset. For nest morphology feature recognition, a convolutional neural network (CNN) architecture is used. Convolutional layers extract the spatial features of the nest, and fully connected layers output morphological feature vectors. For surrounding bird activity feature recognition, a hybrid architecture combining CNN and LSTM is used. First, the CNN extracts the appearance features of a single-frame bird image, then the LSTM processes multiple frames of temporal data to capture the dynamic features of bird activity, outputting an activity feature vector. The morphological feature vector and the activity feature vector are fused according to weights and input into a Softmax classifier to output a bird category probability distribution.

[0047] The labeled dataset is divided into training, validation and test sets for model training and optimization. Nest morphology features and surrounding activity features of the nests to be identified are collected and input into the model to obtain the bird category of the nests to be identified and obtain the stress threshold of noise during the breeding season.

[0048] Furthermore, environmental parameters around the bird's nest were measured, including the terrain attenuation coefficient k and the weather correction factor. vegetation density correction factor The point source spherical wave attenuation model is adopted, and the basic formula is: Where L(r) is the noise value at a distance of r meters from the interference source. To determine the noise source intensity when the interference system outputs electromagnetic waves at the normal interference frequency, adjustments are made according to the different drone models. r is the straight-line distance from the interference source to the target point, and 20lg(r) is the geometric attenuation term for the spherical wave. The normal interference frequency is obtained by calling a preset mapping table between drone models and normal interference frequencies, based on the drone model. The basic formula is then modified to obtain the actual noise propagation formula, specifically including: the downwind noise value in the downwind direction. Headwind noise level In vegetated areas, the noise level of vegetation cover is... When there are multiple types of terrain or vegetation around the bird's nest, the annular area centered on the bird's nest is divided according to the terrain type, and different areas are calculated using the corresponding k value.

[0049] Solve for r such that ,in, This represents the lower limit of the stress threshold, and r is calculated by back-calculating using the actual noise propagation formula; a noise meter is deployed at the calculated r point, and the interference device is turned on until... Measure the actual noise value; when When k is increased, r is recalculated, and the measurement is repeated until... The influence range of all bird nests is summarized to form a set of influence ranges of nest areas.

[0050] In one specific embodiment, a coastal wetland bird breeding ground was used as the experimental setting, and the target birds were the Red-bellied Stint and the Spoon-billed Sandpiper. Sixty nest morphology samples were collected from each bird species (Red-bellied Stint nests were 5-7 cm in diameter, bowl-shaped, and made of dead branches; Spoon-billed Sandpipers nests were 4-6 cm in diameter, disc-shaped, and made of reeds), along with 120 video clips of bird activity from each species (Red-bellied Stints were 23-25 ​​cm long with a grayish-brown back; Spoon-billed Sandpipers were 14-16 cm long with a shovel-shaped beak), forming a labeled dataset (70% training set, 20% validation set, and 10% test set). After model training, the ResNet50 architecture achieved an accuracy of 96% in recognizing bird nest morphology features, while the CNN+LSTM architecture achieved an accuracy of 94% in recognizing bird activity features. Among the 15 bird nests to be identified, the model output 10 for Red-bellied Stints (probability ≥ 92%) and 5 for Spoon-billed Sandpipers (probability ≥ 91%). Matching with the ecological database, the lower limit of the stress threshold during the hatching period for Red-bellied Stints was 44dB, and for Spoon-billed Sandpipers it was 42dB.

[0051] Three typical nesting areas were selected in the experimental breeding site (No. 1 was a mudflat, No. 2 was a reed bed, and No. 3 was a mixed mudflat-reed bed). Environmental parameters were measured: the attenuation coefficient k of the mudflat No. 1 was 0.2 dB / m, and that of the reed bed No. 2 was 0.35 dB / m; the wind speed during the experiment was 4 m / s, the meteorological correction factor f_w in the downwind direction was 0.05 dB / m, and that in the upwind direction was 0.05 dB / m; the vegetation coverage of nesting area No. 1 was 20% (f_v=0), the coverage of nesting area No. 2 was 70% (f_v=0.06 dB / m), and the coverage of the mudflat part of nesting area No. 3 was 15% (f_v=0) and the coverage of the reed bed part was 65% (f_v=0.05 dB / m).

[0052] Taking nesting area 1 (Red-bellied Sandpiper, Tt=44dB) as an example, the interference source is the normal interference frequency (1561MHz) targeting the DJI Mavic 3, with a corresponding noise source intensity L0=83dB. Substituting into the downwind propagation formula for mudflat terrain, the initial calculation yields r=41 meters; deploying a noise meter at r=41 meters, the actual noise value is measured as 45.2dB (>44dB). Increasing k to 0.22dB / m and recalculating, r=39 meters; measuring again, the noise is 43.8dB (≤44dB), determining the radius of influence for nesting area 1 to be 39 meters. Nesting area 2 (Spoon-billed Sandpiper, Tt=42dB) is calculated using the upwind formula for reed beds, ultimately determining the radius of influence to be 32 meters, with an actual measured noise value of 41.7dB.

[0053] Step S200: Scan targets entering bird breeding areas, analyze target features using visual recognition algorithms, match them with the UAV model database, record the UAV model; obtain target coordinates, and update its movement trajectory;

[0054] Specifically, the radar scans bird breeding areas along a preset route, outputs a list of moving targets in real time, sets a threshold for UAV motion characteristics to exclude non-UAV targets, preprocesses the acquired images to obtain a set of feature images to be identified, calls a convolutional neural network algorithm to extract features from the set of feature images to be identified, including structural features, detailed features, and thermal features of the targets, calls a UAV model feature database, uses cosine similarity to calculate the similarity between the feature vector of the target to be identified and the standard feature vector of each model in the database, matches the UAV model and records it.

[0055] In one specific embodiment, the coastal wetland breeding area in step S100 is used as the experimental scenario. An X-band radar (with a scanning range covering the entire breeding area and a 500-meter buffer zone, an azimuth angle of 360°, and a range accuracy of ±3 meters) is deployed and continuously scans along a circular route. During a certain period, the radar outputs 8 moving targets. The drone motion characteristic thresholds are set (flight speed 2-15 m / s, altitude 5-100 meters, smooth trajectory): 2 floating objects with a speed <2 m / s (balloons, fallen leaves), 1 bird with an altitude >100 meters, and 1 insect with a messy trajectory are excluded. The remaining 4 suspected drone targets trigger the camera to perform linked imaging.

[0056] For four suspected targets, images were captured using a 2-megapixel high-definition camera (1920×1080 resolution, 25fps). After Gaussian filtering for noise reduction and contrast enhancement, a set of feature images to be identified was obtained. A ResNet-18 convolutional neural network was then used to extract features: Target 1 has a 3:1 aspect ratio, a quadcopter structure (structural feature vector dimension 256), a "DJI" logo printed on its fuselage (12 detailed key features), and infrared images showing four thermal highlights from its motors (thermal feature vector dimension 64). The corresponding feature vectors for the other three targets were extracted similarly.

[0057] The system utilizes a drone model feature database (containing standard feature vectors for 10 common models including DJI Mavic 3, Xiaomi FIMIX 8, and Haberson Zino Pro) and calculates cosine similarity. Target 1's feature vector shows 93% similarity to the DJI Mavic 3's standard vector, while its similarity to other models is less than 75%. Target 2 shows 91% similarity to the Xiaomi FIMIX 8, and targets 3 and 4 show 90% and 89% similarity to the Haberson Zino Pro, respectively. The four target models are ultimately recorded as DJI Mavic 3, Xiaomi FIMIX 8, Haberson Zino Pro, and Haberson Zino Pro.

[0058] Step S300: Spatial matching of the current target coordinates with the nest area influence range set, and output judgment results; when the target is outside the nest area influence range, generate electromagnetic waves, output frequency is normal interference frequency, and perform navigation interference; when the target is within the nest area influence range, obtain the environmental noise value of the current associated nest area, and determine whether there are birds in the nest corresponding to the current associated nest area through infrared thermal imaging monitoring; determine whether the environmental noise value is greater than the stress threshold and whether the bird is in the nest, and perform sub-scene judgment and interference execution.

[0059] Specifically, the Euclidean distance formula is used to calculate the planar distance between the target and the center point of the nesting area. Based on the set of nesting area influence ranges, it is determined whether the target coordinates fall within the nesting area. When the current nesting area overlaps with other nesting areas, it is simultaneously checked whether the target is in another overlapping nesting area. When the target is in multiple nesting areas at the same time, the target is marked as being in the overlapping area of ​​nesting areas. According to the nesting area distribution from the core area of ​​the breeding ground to the outer area, spatial matching is performed on each valid nesting area in turn, and the judgment result is output.

[0060] Furthermore, based on the current distance between the target and the jamming device, combined with the normal jamming frequency, the output power of the power amplifier is set. According to the target coordinates and the position of the jamming device, the azimuth and elevation angles of the target relative to the jamming device are calculated, and the beam direction of the directional transmitting antenna is adjusted so that the main beam of the antenna is aimed at the target.

[0061] Furthermore, the sensor is activated to continuously collect noise, receiving and storing the environmental noise value every second in real time, and preprocessing the environmental noise value; the infrared thermal imager is activated to collect static thermal images of the bird's nest and surrounding area to determine whether there is a heat source area in the bird's nest area with a temperature higher than the ambient temperature; the dynamic video acquisition mode is switched to determine whether there are dynamic changes in the heat source area and to exclude non-bird heat sources; when there is a heat source area in the static thermal image that matches the characteristics of birds, and bird activity characteristics are observed in the area during dynamic monitoring, it is determined that there are birds in the bird's nest in the currently associated nest area.

[0062] Furthermore, in the first sub-scenario, the ambient noise level is greater than the stress threshold and birds are in their nests. The normal interference frequency corresponding to the target is obtained, and based on the distance between the target and the interference device, the minimum power to suppress the target's navigation system is set. In the second sub-scenario, the ambient noise level is greater than the stress threshold and birds are in their nests. The frequency is set the same as in the first sub-scenario, increasing the power based on the first sub-scenario setting, ensuring no excessive out-of-band radiation and no interference with nearby legitimate devices. In the third sub-scenario, the ambient noise level is not greater than the stress threshold and birds are in their nests. An initial interference frequency is set, lower than the normal interference frequency corresponding to the target. To ensure that the noise generated by the initial interference signal, when superimposed with the current ambient noise, does not exceed the stress threshold, a minimum safe power is set, lower than the power in the first scenario, to prevent the initial power from being too high and causing the noise to exceed the stress threshold. When the target is not moving, the interference frequency is increased by a set ratio. When the superposition still does not exceed the stress threshold, the new frequency is maintained; when it exceeds the threshold, it falls back to the previous frequency and is no longer increased. The increased interference frequency must not exceed the normal interference frequency corresponding to the target. In the fourth sub-scenario, the ambient noise value is not greater than the stress threshold and the birds are not in their nests, so the normal interference frequency corresponding to the target is used, and the power is set to that of the first sub-scenario.

[0063] In one specific embodiment, using the DJI Mavic 3 identified in step S200 as the target (current coordinates: 120.352°E, 30.121°N), the nesting areas of Red-bellied Stint (center coordinates: 120.350°E, 30.120°N, radius of influence: 39 meters) and Spoon-billed Sandpiper (center coordinates: 120.353°E, 30.122°N, radius of influence: 32 meters) are retrieved. Using the Euclidean distance formula, the distance between the target and the center point of nesting area 1 is 28 meters (≤39 meters), and the distance between the target and the center point of nesting area 2 is 21 meters (≤32 meters). Furthermore, there is a 15-meter overlap between the two nesting areas. Therefore, it is determined that "the target is within the overlapping area of ​​nesting areas 1 and 2."

[0064] The jamming device was deployed at coordinates 120.348°E, 30.118°N, with a straight-line distance of 300 meters between the target and the jamming device. Based on the normal jamming frequency of the DJI Mavic 3 (1561MHz) and the distance setting, the initial output power of the power amplifier was set to 22dBm. The azimuth angle of the target relative to the jamming device was calculated to be 18° east of north, and the elevation angle to be 3°. The directional antenna beam was adjusted to this angle to ensure that the main beam covered the target.

[0065] For the associated nesting area No. 1 (stress threshold lower limit 44dB), a noise sensor was deployed to continuously collect the ambient noise value for 30 seconds. After preprocessing, the average noise value was 46dB. Simultaneously, an infrared thermal imager (resolution 640×512) was activated, which detected an irregular elliptical heat source of 39-41℃ in the nesting area (ambient temperature 25℃). After 5 minutes of dynamic monitoring, the heat source was observed to turn its head and make slight standing movements. Non-bird heat sources such as stones were ruled out, and it was determined that there were birds in the nests in nesting area No. 1.

[0066] First sub-scenario (noise 46dB > 44dB, birds present): Use the normal jamming frequency of 1561MHz, set the power to 22dBm (just enough to suppress the target's navigation), and after 10 seconds of jamming, the target begins to move out of the nest area.

[0067] Second sub-scene (noise 46dB > 44dB, birds present): frequency maintained at 1561MHz, power increased to 25dBm (no out-of-band radiation exceeding limits), target movement speed increased from 0.8 m / s to 1.2 m / s;

[0068] Third sub-scene (replaced nest area: nest area 3 noise 42dB≤44dB, with birds): initial frequency 1171MHz (normal frequency 75%), power 13dBm, after the target does not move, increase to 1239MHz by 5%, superimposed noise 43.5dB≤44dB, maintain this frequency;

[0069] Fourth sub-scenario (Noise level of nest area 3: 42dB≤44dB, no birds): Using the normal frequency of 1561MHz and the power of 22dBm, the target moves towards the boundary point after 8 seconds of interference.

[0070] Step S400: During the navigation interference process, filter out points that deviate from the boundary and generate a predetermined trajectory; track the target trajectory and perform navigation interference on the target based on the predetermined trajectory. When the target's current coordinates deviate from the bird breeding area, the navigation interference terminates.

[0071] Specifically, the selection criteria for points that have deviated from the boundary are as follows, in descending order of priority: points that are on the boundary of the breeding ground; the straight-line distance from the point to the safe landing zone is relatively shorter among all boundary points; the line connecting the target's current location to the point passes through the fewest nesting areas; and there are no ecologically sensitive areas in or around the point.

[0072] The boundary line of the breeding ground is decomposed into several continuous boundary points, which are then selected according to screening criteria to obtain the boundary-free points. The target coordinates are used as the starting point of the predetermined trajectory, and the boundary-free points are used as the ending point of the predetermined trajectory. The starting point and the ending point of the trajectory are connected to form an initial straight-line trajectory. Trajectory constraints are set, including: the entire trajectory does not enter any nesting area; the deviation between the total trajectory length and the straight-line distance from the starting point to the ending point is minimized; the trajectory does not cross ecologically sensitive areas or the protection range of legal equipment.

[0073] When the initial straight trajectory fully conforms to the trajectory constraints, it is taken as the predetermined trajectory. When the initial straight trajectory crosses any region in the trajectory constraints, the path inflection point that bypasses the region with the least increase in distance is found on both sides of the crossed region to form the initial predetermined trajectory. The inflection point of the initial predetermined trajectory is smoothed. Based on the target's flight capability, it is determined whether the smoothed initial predetermined trajectory is within the target's executable range. If it is outside the range, the position of the trajectory inflection point is further adjusted until the trajectory conforms to the target's flight capability and the predetermined trajectory is formed.

[0074] In one specific embodiment, a coastal wetland breeding ground is used as the scenario. The boundary of the breeding ground is a polygon (including a 500-meter buffer zone), and the safe drop zone is the open mudflat on the east side of the boundary (no nesting area, no sensitive area). The boundary line of the breeding ground is decomposed into 50 continuous boundary points, and filtered according to the priority of the screening criteria: 20 points with "straight-line distance to the safe drop zone < 80 meters" are retained first, and then 8 points are selected from them with "the line connecting the target's current coordinates (120.352°E, 30.121°N) to this point only crosses 1 nesting area". Finally, the boundary point (120.358°E, 30.123°N) with "no ecologically sensitive area within 10 meters" is determined as the departure point from the boundary. The distance from this point to the safe drop zone is 72 meters, and the number of nesting areas crossed by the line is 1 (only the edge of the No. 2 Spoon-billed Sandpiper nesting area).

[0075] Starting from the target's current coordinates (120.352°E, 30.121°N) and ending at the aforementioned boundary point, an initial straight trajectory is formed, approximately 110 meters in length. Trajectory constraints are checked: the trajectory crosses the influence area of ​​nest zone 2 (radius 32 meters) in the middle section (at 120.355°E, 30.122°N) and is close to the chicks' sensitive activity area (5 meters away), violating the constraint of "not entering nest zone and sensitive area," thus requiring trajectory adjustment.

[0076] Inflection points were set on both sides of the initial trajectory passing through Nest Area 2: Inflection Point 1 (120.354°E, 30.120°N) and Inflection Point 2 (120.356°E, 30.120°N), forming an initial predetermined trajectory of "start point - inflection point 1 - inflection point 2 - end point", with a total length of 125 meters (13.6% deviation from the original straight line distance). The inflection points were smoothed: the broken line between inflection point 1 and inflection point 2 was adjusted into a gentle curve (the turning angle was reduced from the original 60° to 35°). Combined with the DJI Mavic 3's flight capability of a maximum turning angle of 40°, the adjusted trajectory was verified to be within the target's executable range. The final predetermined trajectory had a total length of 128 meters, with the deviation controlled within 16.4%.

[0077] After initiating navigation interference, the target coordinates were updated every 5 seconds: after 10 seconds of interference, the target reached inflection point 1 (120.354°E, 30.120°N); after 20 seconds, it reached inflection point 2 (120.356°E, 30.120°N); and after 35 seconds, it reached the boundary departure point (120.358°E, 30.123°N). Monitoring continued for 5 seconds, confirming the target coordinates were (120.359°E, 30.124°N), indicating it had left the breeding ground boundary (8 meters from the boundary) and was moving towards a safe landing zone. Interference was immediately terminated. The target moved along the predetermined trajectory without deviation throughout the entire process.

[0078] like Figure 2 As shown, this application provides a deflection system for targets whose navigation is interfered with by electromagnetic waves of a specific frequency, comprising:

[0079] Nesting Area Impact Range Generation Module: Includes: Nest Coordinate Positioning Unit, Stress Threshold Acquisition Unit, and Nesting Area Impact Range Generation Unit; wherein, the Nest Coordinate Positioning Unit acquires the three-dimensional coordinates of all nests within the bird breeding grounds, forming a nest coordinate set; the Stress Threshold Acquisition Unit uses AI image recognition algorithms to identify the bird species to which each nest belongs and acquires its stress threshold for noise during the breeding season; the Nesting Area Impact Range Generation Unit uses each nest coordinate as the center point and, based on the stress threshold, calculates the noise impact range from the inside out, forming a nesting area impact range set;

[0080] Target analysis module: includes a drone model acquisition unit and a target coordinate acquisition unit; wherein, the drone model acquisition unit scans targets entering the bird breeding area, analyzes the target features using a visual recognition algorithm, matches them with the drone model database, and records the drone model; the target coordinate acquisition unit acquires the target coordinates and updates its movement trajectory;

[0081] The navigation interference execution module includes: a spatial matching unit, an out-of-range interference execution unit, an in-range interference data acquisition unit, and an in-range interference execution unit. The spatial matching unit performs spatial matching between the current target coordinates and the nest area influence range set, outputting the judgment result. The out-of-range interference execution unit generates electromagnetic waves when the target is outside the nest area influence range, outputting the normal interference frequency for navigation interference. The in-range interference data acquisition unit acquires the environmental noise value of the currently associated nest area when the target is within the nest area influence range, and uses infrared thermal imaging to determine whether there are birds in the corresponding nests. The in-range interference execution unit determines whether the environmental noise value exceeds the stress threshold and whether birds are in the nests, performing sub-scene judgment and interference execution.

[0082] The predetermined trajectory generation and execution module includes a predetermined trajectory generation unit and a predetermined trajectory execution unit. The predetermined trajectory generation unit filters out points that deviate from the boundary during the navigation interference process and generates a predetermined trajectory. The predetermined trajectory execution unit tracks the target trajectory and performs navigation interference on the target based on the predetermined trajectory. When the target's current coordinates deviate from the bird breeding area, the navigation interference terminates.

[0083] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for driving away targets whose navigation is interfered with by electromagnetic waves of a specific frequency, characterized in that, Includes the following steps: The three-dimensional coordinates of all bird nests within the bird breeding grounds are obtained to form a set of bird nest coordinates; the bird species to which each nest belongs is identified through AI image recognition algorithms, and its stress threshold for noise during the breeding season is obtained; with each bird nest coordinate as the center point, the noise impact range from the inside out is calculated based on the stress threshold, forming a set of nest area impact ranges. The method uses the coordinates of each bird's nest as the center point and, based on the stress threshold, calculates the noise impact range from the inside out, forming a set of nest area impact ranges, including: Measure environmental parameters around the bird's nest, including the terrain attenuation coefficient k and the weather correction factor. vegetation density correction factor The point source spherical wave attenuation model is adopted, and the basic formula is: Where L(r) is the noise value at a distance of r meters from the interference source. To determine the noise source intensity when the interference system outputs electromagnetic waves at the normal interference frequency, adjustments are made according to the different drone models. r is the straight-line distance from the interference source to the target point, and 20lg(r) is the geometric attenuation term for the spherical wave. The normal interference frequency is obtained by calling a preset mapping table between drone models and normal interference frequencies, based on the drone model. The basic formula is then modified to obtain the actual noise propagation formula, specifically including: the downwind noise value in the downwind direction. Headwind noise level In vegetated areas, the noise level of vegetation cover is... When there are multiple types of terrain or vegetation around the bird's nest, the annular area centered on the bird's nest is divided according to the terrain type, and different areas are calculated using the corresponding k value. Solve for r such that ,in, This represents the lower limit of the stress threshold, and r is calculated by back-calculating using the actual noise propagation formula; a noise meter is deployed at the calculated r point, and the interference device is turned on until... Measure the actual noise value; when When k is increased, r is recalculated, and the measurement is repeated until... ; Summarize the influence range of all bird nests to form a set of influence ranges of nest areas; The system scans targets entering bird breeding areas, analyzes their features using visual recognition algorithms, matches them against a drone model database, records the drone's model, obtains the target's coordinates, and updates its trajectory. Spatial matching is performed between the current target coordinates and the set of nest area influence ranges, and the judgment result is output. When the target is outside the nest area influence range, electromagnetic waves are generated, and the output frequency is the normal interference frequency to perform navigation interference. When the target is within the nest area influence range, the environmental noise value of the current associated nest area is obtained, and infrared thermal imaging is used to determine whether there are birds in the nests corresponding to the current associated nest area. The environmental noise value is judged to be greater than the stress threshold and whether there are birds in the nests, and sub-scene judgment and interference execution are performed. During the navigation interference process, points that deviate from the boundary are selected and a predetermined trajectory is generated; the target trajectory is tracked and navigation interference is carried out on the target based on the predetermined trajectory; the navigation interference terminates when the target's current coordinates leave the bird breeding area. The process of filtering out boundary points and generating a predetermined trajectory during navigation interference includes: The selection criteria for points that have deviated from the boundary are ranked from highest to lowest as follows: points that are on the boundary of the breeding grounds; the straight-line distance from the point to the safe landing zone is relatively shorter among all boundary points; the line connecting the target's current location to the point passes through the fewest nesting areas; and there are no ecologically sensitive areas in or around the point. The boundary line of the breeding ground is decomposed into several continuous boundary points, which are then selected according to screening criteria to obtain the boundary-free points. The target coordinates are used as the starting point of the predetermined trajectory, and the boundary-free points are used as the ending point of the predetermined trajectory. The starting point and the ending point of the trajectory are connected to form an initial straight-line trajectory. Trajectory constraints are set, including: the entire trajectory does not enter any nesting area; the deviation between the total trajectory length and the straight-line distance from the starting point to the ending point is minimized; the trajectory does not cross ecologically sensitive areas or the protection range of legal equipment. When the initial straight trajectory fully conforms to the trajectory constraints, it is taken as the predetermined trajectory. When the initial straight trajectory crosses any region in the trajectory constraints, the path inflection point that bypasses the region with the least increase in distance is found on both sides of the crossed region to form the initial predetermined trajectory. The inflection point of the initial predetermined trajectory is smoothed. Based on the target's flight capability, it is determined whether the smoothed initial predetermined trajectory is within the target's executable range. If it is outside the range, the position of the trajectory inflection point is further adjusted until the trajectory conforms to the target's flight capability and the predetermined trajectory is formed.

2. The method for driving away navigation targets that are interfered with by electromagnetic waves of a specific frequency, as described in claim 1, is characterized in that... The process of using AI image recognition algorithms to identify the bird species to which each nest belongs and obtaining its stress threshold for noise during the breeding season includes: This study collects nest morphology features and surrounding activity characteristics of potential bird species in bird breeding grounds. Nest morphology features include nest size, shape, material, and nesting location characteristics. Surrounding activity characteristics include bird size, feather color, and behavioral patterns. Each nest morphology image and each video frame of bird activity is labeled with a corresponding bird category label, and key feature regions are marked to form a labeled dataset. For nest morphology feature recognition, a convolutional neural network (CNN) architecture is used. Convolutional layers extract spatial features of the nest, and fully connected layers output morphological feature vectors. For surrounding bird activity feature recognition, a hybrid architecture combining CNN and LSTM is used. First, the CNN extracts the appearance features of a single-frame bird image, then the LSTM processes multiple frames of temporal data to capture the dynamic features of bird activity, outputting an activity feature vector. The morphological feature vector and the activity feature vector are fused according to weights and input into a Softmax classifier to output a bird category probability distribution. The labeled dataset is divided into training, validation and test sets for model training and optimization. Nest morphology features and surrounding activity features of the nests to be identified are collected and input into the model to obtain the bird category of the nests to be identified and obtain the stress threshold of noise during the breeding season.

3. The method for driving away navigation targets that are interfered with by electromagnetic waves of a specific frequency, as described in claim 1, is characterized in that... The scanning of targets entering bird breeding areas employs visual recognition algorithms to analyze target features, matches them against a drone model database, and records the drone model, including: The radar scans the bird breeding area along a preset route, outputs a list of moving targets in real time, sets a threshold for UAV motion characteristics to exclude non-UAV targets, preprocesses the acquired images to obtain a set of feature images to be identified, calls a convolutional neural network algorithm to extract features from the set of feature images to be identified, including structural features, detailed features, and thermal features of the targets, calls a UAV model feature database, uses cosine similarity to calculate the similarity between the feature vector of the target to be identified and the standard feature vector of each model in the database, matches the UAV model and records it.

4. The method for driving away navigation targets that are interfered with by electromagnetic waves of a specific frequency, as described in claim 1, is characterized in that... The step of spatially matching the current target coordinates with the set of influence areas of the nest region and outputting the judgment result includes: The Euclidean distance formula is used to calculate the planar distance between the target and the center point of the nest area. Based on the set of influence ranges of the nest area, it is determined whether the target coordinates fall within the nest area. When the current nest area overlaps with other nest areas, it is simultaneously checked whether the target is in another overlapping nest area. When the target is in the range of multiple nest areas at the same time, the target is marked as being in the overlapping area of ​​nest areas. According to the distribution of nest areas from the core area of ​​the breeding ground to the outer area, spatial matching is performed on each valid nest area in turn, and the judgment result is output.

5. The method for driving away a target whose navigation is interfered with by electromagnetic waves of a specific frequency, as described in claim 1, is characterized in that... When the target is outside the influence range of the nest area, electromagnetic waves are generated, with the output frequency being the normal interference frequency, to perform navigation interference, including: Based on the current distance between the target and the jamming device, combined with the normal jamming frequency, the output power of the power amplifier is set. According to the target coordinates and the position of the jamming device, the azimuth and elevation angles of the target relative to the jamming device are calculated, and the beam direction of the directional transmitting antenna is adjusted so that the main beam of the antenna is aimed at the target.

6. The method for driving away a target whose navigation is interfered with by electromagnetic waves of a specific frequency, as described in claim 1, is characterized in that... When the target is within the influence range of the nesting area, the environmental noise value of the currently associated nesting area is obtained, and infrared thermal imaging is used to determine whether there are birds in the nests corresponding to the currently associated nesting area, including: The sensor is activated to continuously collect noise data, receiving and storing the environmental noise value every second in real time, and preprocessing the environmental noise value; the infrared thermal imager is activated to collect static thermal images of the bird's nest and surrounding area to determine whether there is a heat source area in the bird's nest area with a temperature higher than the ambient temperature; the dynamic video acquisition mode is switched to determine whether there are dynamic changes in the heat source area and to exclude non-bird heat sources; when there is a heat source area in the static thermal image that matches the characteristics of birds, and bird activity characteristics are observed in the area during dynamic monitoring, it is determined that there are birds in the bird's nest in the currently associated nest area.

7. The method for driving away navigation targets that are interfered with by electromagnetic waves of a specific frequency, as described in claim 1, is characterized in that... The process of determining whether the environmental noise level exceeds the stress threshold and whether the bird is in its nest, and performing sub-scene judgment and interference execution, includes: In the first sub-scenario, the ambient noise level exceeds the stress threshold and birds are in their nests. The normal interference frequency corresponding to the target is obtained, and the minimum power to suppress the target's navigation system is set based on the distance between the target and the interfering device. In the second sub-scenario, the ambient noise level exceeds the stress threshold and birds are in their nests. The frequency is set the same as in the first sub-scenario, increasing the power based on the first sub-scenario setting, ensuring no excessive out-of-band radiation and no interference with nearby legitimate devices. In the third sub-scenario, the ambient noise level is not greater than the stress threshold and birds are in their nests. An initial interference frequency is set, lower than the normal interference frequency corresponding to the target, to ensure initial... If the noise generated by the interference signal, when superimposed with the current ambient noise, still does not exceed the stress threshold, a minimum safe power is set, lower than the power of the first scenario, to avoid the initial power being too high and causing the noise to exceed the stress threshold. When the target is not moving, the interference frequency is increased by a set ratio. If the superimposed noise still does not exceed the stress threshold, the new frequency is maintained. If it exceeds the threshold, it falls back to the previous frequency and is no longer increased. The increased interference frequency must not exceed the normal interference frequency corresponding to the target. In the fourth sub-scenario, the ambient noise value is not greater than the stress threshold and the birds are not in their nests. The normal interference frequency corresponding to the target is used, and the power is set to that of the first sub-scenario.

8. A system for driving away targets whose navigation is interfered with by electromagnetic waves of a specific frequency, using the method for driving away targets whose navigation is interfered with by electromagnetic waves of a specific frequency as described in any one of claims 1-7, characterized in that, include: Nesting Area Impact Range Generation Module: Includes: Nest Coordinate Positioning Unit, Stress Threshold Acquisition Unit, and Nesting Area Impact Range Generation Unit; wherein, the Nest Coordinate Positioning Unit acquires the three-dimensional coordinates of all nests within the bird breeding grounds, forming a nest coordinate set; the Stress Threshold Acquisition Unit uses AI image recognition algorithms to identify the bird species to which each nest belongs and acquires its stress threshold for noise during the breeding season; the Nesting Area Impact Range Generation Unit uses each nest coordinate as the center point and, based on the stress threshold, calculates the noise impact range from the inside out, forming a nesting area impact range set; Target analysis module: includes a drone model acquisition unit and a target coordinate acquisition unit; wherein, the drone model acquisition unit scans targets entering the bird breeding area, analyzes the target features using a visual recognition algorithm, matches them with the drone model database, and records the drone model; the target coordinate acquisition unit acquires the target coordinates and updates its movement trajectory; The navigation interference execution module includes: a spatial matching unit, an out-of-range interference execution unit, an in-range interference data acquisition unit, and an in-range interference execution unit. The spatial matching unit performs spatial matching between the current target coordinates and the nest area influence range set, outputting the judgment result. The out-of-range interference execution unit generates electromagnetic waves when the target is outside the nest area influence range, outputting the normal interference frequency for navigation interference. The in-range interference data acquisition unit acquires the environmental noise value of the currently associated nest area when the target is within the nest area influence range, and uses infrared thermal imaging to determine whether there are birds in the corresponding nests. The in-range interference execution unit determines whether the environmental noise value exceeds the stress threshold and whether birds are in the nests, performing sub-scene judgment and interference execution. The predetermined trajectory generation and execution module includes a predetermined trajectory generation unit and a predetermined trajectory execution unit. The predetermined trajectory generation unit filters out points that deviate from the boundary during the navigation interference process and generates a predetermined trajectory. The predetermined trajectory execution unit tracks the target trajectory and performs navigation interference on the target based on the predetermined trajectory. When the target's current coordinates deviate from the bird breeding area, the navigation interference terminates.

Images