Method for autonomous interception of aerial drones based on ground camera network

Abstract

The application relates to the technical field of unmanned aerial vehicle interception, and discloses an aerial unmanned aerial vehicle autonomous interception method based on a ground camera network, which comprises the following steps: a node camera in the ground camera network monitors a no-fly area and continuously collects images; an intrusion target object in the image collected by each node camera is detected through a target detection algorithm, and the two-dimensional coordinates of the intrusion target object in the imaging plane of the node camera are calculated; the position of the intrusion target object in a three-dimensional space is solved; an interception trajectory of an interception unmanned aerial vehicle is designed; and a trajectory tracking controller of the interception unmanned aerial vehicle is designed to control the interception unmanned aerial vehicle to intercept the intrusion target object. The application can avoid the occurrence of target loss caused by the self maneuvering of the interception unmanned aerial vehicle and the escape flight of the intrusion target object, can reduce the hardware load of the interception unmanned aerial vehicle, can reduce the operation burden of the interception unmanned aerial vehicle, and can improve the continuous tracking and interception capacity of the interception unmanned aerial vehicle on the intrusion target object.

Description

Technical Field

[0001] This invention relates to the field of drone interception technology, and specifically to an autonomous interception method for aerial drones based on a ground camera network. Background Technology

[0002] In recent years, with the continuous improvement of the safety, flexibility, and operability of drones, their applications are no longer limited to research institutes and specific industries. The widespread adoption of commercial drones has led to a year-on-year increase in their ownership among the public. How to effectively drive away and intercept small drones that intrude into no-fly zones has become a hot research topic. Existing counter-drone technologies mainly include using radio wave guns to intercept drone remote control signals, using GPS jammers to disrupt drone positioning information, and using interception nets for physical interception. However, because counter-drone equipment generally has a limited range of effectiveness, and interfering with and physically intercepting intruding drones can easily cause them to crash and cause secondary damage, using a drone to autonomously approach and intercept intruding drones is a more feasible and efficient solution.

[0003] Due to limitations in the payload and computing power of drones, adding high-precision sensors and running complex image processing algorithms will inevitably affect the drone's endurance. Furthermore, due to the inherent hardware characteristics of airborne vision cameras, they can only effectively perceive intruding targets within a specific direction and distance range. Therefore, when an intruding target suddenly accelerates or the intercepting drone makes significant attitude adjustments, the target image or visual features used to provide control error signals are easily lost. Existing autonomous aerial drone interception systems based on visual feedback signals must rely on target visual information to continuously adjust the intercepting drone's flight state; thus, target or feature loss inevitably leads to interception mission failure. In other words, autonomous flight interception methods using drone-borne vision sensors cannot fundamentally solve the problem of target feature loss caused by the camera's field of view limitations.

[0004] Currently, research on autonomous flight systems for unmanned aerial vehicles (UAVs) based on ground-based camera networks is relatively limited. Early work on camera networks primarily focused on exploring scene coverage. This involved optimization methods to achieve full surveillance coverage of a fixed scene using the fewest possible camera nodes, and optimizing their positions and orientations. More recent work has focused on distributed control of each rotatable camera node within the camera network to continuously track multiple freely moving intruding targets within the monitored scene. However, existing work on camera networks focuses solely on optimizing their own surveillance capabilities, neglecting the significant potential of their wide-field-of-view perception for robot localization and navigation. While some recent research has explored aerial UAV localization and trajectory prediction based on sensor networks, the algorithmic performance relies on specially customized combinations of sensor nodes, rendering these methods unsuitable for many practical scenarios with only standard surveillance cameras.

[0005] In the camera network-based autonomous UAV interception and control method, the camera network is a sensor network composed of cameras as the main nodes. By optimizing the field of view of each camera in the network, it can achieve full-coverage monitoring of a specific area. For the task of autonomously intercepting intruding UAVs in no-fly zones, this invention uses a ground-based camera network as the system's perception module, aiming to propose an autonomous UAV interception method and system based on a ground-based camera network. This not only effectively utilizes existing video surveillance systems in special areas such as airports and schools, relying on their wide field of view to monitor and track intruding aircraft throughout their flight, avoiding target loss due to the interceptor UAV's own maneuvering and the intruding target's escape flight, but also reduces the hardware load on the interceptor UAV, lowers its computational burden, and effectively improves its continuous tracking and interception capabilities against intruding targets. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides an autonomous interception method for aerial unmanned aerial vehicles (UAVs) based on a ground-based camera network. Targeting the mission context of tracking and intercepting intruding malicious UAVs in no-fly zones such as airports and schools, this method leverages existing camera monitoring networks in these areas. It estimates the position of the intruding UAV through techniques such as moving target detection, image feature extraction, and matching. An autonomous flight controller is designed based on the relative position information of the intruding target and the intercepting UAV, aiming to achieve effective autonomous interception of the intruding target by the intercepting UAV. The visual perception module of this method no longer relies on the onboard visual sensors of the intercepting UAV but instead utilizes target information provided by the ground-based camera network. This provides an effective solution to the problem of target feature loss in research on vision-based autonomous interception of aerial UAVs.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0008] An autonomous interception method for aerial drones based on a ground camera network includes the following steps:

[0009] Step 1: Node cameras in the ground camera network monitor the no-fly zone and continuously collect images;

[0010] Step 2: Using a target detection algorithm, detect intrusion targets in the images captured by each node camera and calculate the two-dimensional coordinates of the intrusion targets on the imaging plane of the node camera.

[0011] Step 3: Record the number N of node cameras that detected the same intrusion target;

[0012] Step 4: Determine if N is greater than or equal to 2; if yes, proceed to Step 5; otherwise, return to Step 1.

[0013] Step 5: Based on the distribution and orientation information of the node cameras in the ground camera network, calculate the position of the intrusion target in three-dimensional space using the two-dimensional coordinates of the intrusion target on the imaging plane of different node cameras in the ground camera network.

[0014] Step 6: Based on the relative three-dimensional spatial positions of the intrusion target and the intercepting drone, design the interception trajectory of the intercepting drone;

[0015] Step 7: Design an interceptor drone trajectory tracking controller to track the interception trajectory and control the interceptor drone to intercept the intruding target.

[0016] Furthermore, step two specifically includes the following steps:

[0017] S21. In the absence of intrusive targets, establish background models for each node camera in the ground camera network, where the background model established for the i-th node camera is represented by B. i It means that B i (u,n) represents the pixel grayscale value in the u-th row and n-th column of the background model of the i-th node camera;

[0018] S22. Each node camera continuously acquires images and calculates the pixel grayscale difference D. i (u,n)=|I i (u,n)-B i (u,n)|, where I i (u,n) represents the pixel grayscale value in the u-th row and n-th column of the image currently captured by the i-th node camera;

[0019] S23, Determine D i If (u,n) is greater than the threshold T, then it is I. i(u,n) is assigned the value 255, otherwise it is I. i (u,n) is assigned the value 0;

[0020] S24. Perform opening and dilation operations in the image morphology operations in sequence;

[0021] S25. Generate the smallest bounding box that encloses the connected components in the image, and output the center pixel coordinates (n) of the smallest bounding box. t ,n t This gives the two-dimensional coordinates of the intruding target on the imaging plane of the node camera.

[0022] Furthermore, step five specifically includes the following steps:

[0023] S51. Calculate the horizontal deflection angle ψ of the projection of the intruding target onto the imaging plane of the camera at the i-th node. i1 The vertical deflection angle θ i1 :

[0024] ψ i1 =atan2(λ,(u t -u0)du),θ i1 =atan2(λ / cosψ i1 ,(n t -n0)dn);

[0025] Where atan2 represents the function for calculating the azimuth angle, λ represents the focal length of the nodal camera, (u0, n0) represents the center pixel coordinates of the acquired image, and du and dn represent the length and width of a pixel in the acquired image, respectively. t ,n t () represents the two-dimensional coordinates of the intrusion target on the imaging plane of the node camera;

[0026] S52. Calculate the overall deflection angle ψ of the intruding target relative to the i-th node camera. i and θ i :

[0027] ψ i =ψ i1 +ψ i2 ,θ i =θ i1 +θ i2 ;

[0028] Where ψ i2 and θ i2 These represent the horizontal and vertical deflection angles of the camera at the i-th node, respectively.

[0029] S53. Repeat steps S51 and S52 to calculate the overall deflection angle ψ of the intruding target relative to the camera at the j-th node. j and θj ;

[0030] S54. Calculate the position coordinates (x, y) of the intruding target in three-dimensional space. ij ,y ij ,z ij ):

[0031] y ij =D / (tanψ) i -tanψ j ),x ij =y ij tanψ i ,z ij =y ij tanθ i / cosψ i ;

[0032] Where D represents the distance between the i-th node camera and the j-th node camera;

[0033] S55. If N equals 2 in step four, then the actual position coordinates of the intruding target in three-dimensional space are (x... ij ,y ij ,z ij If N is greater than 2, the node cameras that detect the intrusion target in the image are divided into pairs and the position coordinates of the intrusion target in the three-dimensional space corresponding to each group of node cameras are calculated according to steps S51 to S54. The actual position coordinates of the intrusion target in the three-dimensional space are taken as the average of the position coordinates of the intrusion target in the three-dimensional space corresponding to each group of node cameras.

[0034] Furthermore, step six specifically includes the following steps:

[0035] S61. Design an interception trajectory for intercepting drones. d (t):

[0036]

[0037] Where a k represents the coefficients of the k-th order polynomial, and T represents the time interval for flight trajectory planning;

[0038] S62. Determine the constraints: where the initial position S d (0) represents the three-dimensional spatial coordinates of the intercepting drone, with the termination position S. d (T) represents the three-dimensional spatial position of the intrusion target. The velocity and acceleration constraints at the start and end times are determined according to the actual scenario.

[0039] S63. Based on the constraints, solve for the polynomial coefficients in the interception trajectory to determine the interception trajectory.

[0040] Furthermore, step seven specifically includes:

[0041] S71, based on the desired interception trajectory S of the intercepting drone. d Define the trajectory tracking error δ = SS d , where S is the actual location of the intercepted drone;

[0042] S72, Design of a trajectory tracking controller for intercepting drones:

[0043] u=k p δ+k d dδ;

[0044] Where u represents the control input, k p and k d All of these indicate adjustable control parameters.

[0045] Furthermore, the interception system employed includes:

[0046] The ground sensing unit, which is connected to the communication transmission unit, consists of cameras fixedly installed on the ground or buildings. Each camera acts as a node camera and is connected to other node cameras through data transmission lines, thus forming a ground camera network. Data from each node camera in the ground camera network is transmitted to the central computing unit through the communication transmission unit.

[0047] The central computing unit, connected to the communication transmission unit, receives images captured by node cameras in the ground sensing unit, runs target detection algorithms to perceive intruding targets in the images captured by the node cameras, runs spatial positioning algorithms to calculate the three-dimensional spatial coordinates of the intruding target, runs trajectory planning algorithms to design the interception trajectory of the UAV based on the relative positional relationship between the intercepting UAV and the intruding target, runs tracking control algorithms to track the flight trajectory of the intercepting UAV, and finally, the control signals obtained by the tracking control algorithms are transmitted to the interception execution unit through the communication transmission unit.

[0048] The communication transmission unit includes a wired transmission module and a wireless transmission module. The wired transmission module connects the ground sensing unit and the central computing unit through a data transmission line, transmitting the visual image data acquired by the ground sensing unit to the central computing unit for intrusion target detection. The wireless transmission module connects the central computing unit and the execution interception unit through radio signals, transmitting the trajectory tracking control signals generated by the central computing unit to the execution interception unit for interception flight.

[0049] The interception unit, i.e. the interceptor drone entity, is connected to the communication transmission unit. Through the airborne flight controller, it converts the trajectory tracking control signal generated by the central computing unit into the rotation speed of the interceptor drone's actuators, thereby achieving the tracking and interception of the intruding target.

[0050] Compared with the prior art, the beneficial technical effects of the present invention are:

[0051] This invention aims to intercept low, slow, and small drones that intrude into no-fly zones using a ground-based visual sensor network. Without requiring its own sensing module, the system calculates the state information of the intruding drone through the sensor network, and uses this information, along with surrounding environmental information, to plan the interceptor's trajectory. A controller is designed to efficiently track the flight trajectory, ultimately achieving an automated aerial drone interception system. Specifically, this invention changes the existing aerial drone interception methods and systems that rely on airborne sensors for target perception. Instead, it uses a ground-based camera network for effective perception, localization, and tracking of intruding drones. First, a background description model is designed based on the characteristics of the actual scene and the intruding target, and the background subtraction method is used to effectively perceive the intruding drone. Second, an aerial positioning algorithm for the intruding target is designed based on the distribution and orientation information of camera nodes in the network. The drone's position in three-dimensional space is calculated using the two-dimensional coordinates of the intruding drone on different camera imaging planes in the network. Subsequently, an interception trajectory is planned based on the relative position information of the intruding drone and the interceptor drone. Finally, a trajectory tracking controller for the interceptor drone is designed to achieve stable tracking of the interception trajectory, thereby successfully intercepting the intruding target. Attached Figure Description

[0052] Figure 1 A schematic diagram of the composition of the aerial unmanned aerial vehicle autonomous interception system based on a ground camera network in this invention;

[0053] Figure 2 This is a schematic diagram of the algorithm used in the autonomous interception method for aerial drones based on a ground camera network in this invention.

[0054] Figure 3 This is a flowchart illustrating the autonomous interception method for aerial drones based on a ground camera network in this invention. Detailed Implementation

[0055] A preferred embodiment of the present invention will now be described in detail with reference to the accompanying drawings.

[0056] The autonomous aerial drone interception system based on a ground camera network of the present invention includes: a ground sensing unit, a central computing unit, a communication transmission unit, and an interception execution unit. The connection relationships between the units are as follows: Figure 1 As shown.

[0057] The ground sensing unit, which is connected to the communication transmission unit, consists of cameras fixedly installed on the ground or buildings. Each camera acts as a node and is connected to other camera nodes through data transmission lines to form a ground camera network. At the same time, the data of each camera node in the ground camera network is transmitted to the central computing unit through the communication transmission unit.

[0058] The central computing unit, connected to the communication transmission unit, receives visual image data from the node cameras in the ground sensing unit, runs target detection algorithms to effectively perceive intruding targets in the node camera images, runs spatial positioning algorithms to calculate the three-dimensional spatial coordinates of the intruding targets, runs trajectory planning algorithms to design efficient interception trajectories for the UAV based on the relative positional relationship between the intercepting UAV and the intruding targets, runs tracking control algorithms to achieve stable tracking of the intercepting UAV's flight trajectory, and finally, the control signals obtained by the tracking control algorithms are transmitted to the interception execution unit through the communication transmission unit.

[0059] The communication transmission unit is divided into a wired transmission module and a wireless transmission module. The wired transmission module connects the ground sensing unit and the central computing unit through a data transmission line, transmitting the visual image data acquired by the ground sensing unit to the central computing unit for intrusion target detection. The wireless transmission module connects the central computing unit and the execution interception unit through radio signals, transmitting the control signals generated by the central computing unit to the execution interception unit for interception.

[0060] The interception unit, connected to the communication transmission unit, intercepts the drone entity. Through the onboard flight controller, it converts the trajectory tracking control signal generated by the central computing unit into the rotation speed of the drone's actuators, thereby achieving continuous tracking and interception of the intruding target.

[0061] like Figure 2 As shown, the autonomous interception method for aerial drones based on a ground camera network of the present invention employs the following algorithm:

[0062] The target detection algorithm detects whether the images of each node camera in the ground sensing unit, i.e. the ground camera network, contain an intruding target. For node cameras containing images of intruding targets, the algorithm calculates the two-dimensional coordinate information of the intruding target on the imaging plane.

[0063] The spatial positioning algorithm calculates the position information of the intrusion target in three-dimensional space by using the two-dimensional coordinates of the imaging plane of different node cameras in the network, based on the distribution and orientation information of the node cameras in the ground camera network.

[0064] The trajectory planning algorithm designs an efficient interception trajectory for the drone based on the relative three-dimensional spatial position information of the intruding target and the intercepting drone.

[0065] A tracking control algorithm was developed, and a trajectory tracking controller for intercepting drones was designed to achieve stable tracking of the interception trajectory.

[0066] The specific process of the autonomous interception method for aerial drones based on a ground camera network of the present invention is as follows: Figure 3 As shown, it includes the following steps:

[0067] Step 1: The ground sensing unit, i.e., the node cameras in the ground camera network, monitors the no-fly zone, continuously collects environmental images, and transmits the visual image data to the central computing unit through the communication transmission unit.

[0068] Step 2: The central computing unit runs the target detection algorithm to detect intrusion targets in the images captured by each node camera in the ground sensing unit, and calculates and transmits the two-dimensional coordinate information of the intrusion targets on the imaging plane.

[0069] Step 3: Record the number N of node cameras that detected the intrusion target in the acquired images.

[0070] Step 4: Determine whether the number N of the node cameras that detected the intrusion target in the acquired image is greater than or equal to 2. If yes, continue to step 5; otherwise, return to step 1.

[0071] Step 5: Based on the distribution and orientation information of the node cameras in the ground camera network, calculate the position information of the intrusion target in three-dimensional space by using the two-dimensional coordinates of the imaging plane of different node cameras in the network.

[0072] Step 6: Based on the relative three-dimensional spatial position information of the intrusion target and the intercepting drone, design an efficient interception trajectory for the intercepting drone.

[0073] Step 7: Design an interceptor drone trajectory tracking controller to achieve stable tracking of the interception trajectory, thereby successfully intercepting the intruding target.

[0074] Step two specifically includes the following steps:

[0075] S21. In the absence of intrusive targets, each node camera in the ground sensing unit, i.e., the ground camera network, establishes a background model, where the background model established by the i-th node camera is represented by B. i It means that B i (u,n) represents the pixel grayscale value in the u-th row and n-th column of the camera background model of this node.

[0076] S22. Each node camera continuously acquires images and calculates D. i (u,n)=|I i (u,n)-B i (u,n)|, where Ii (u,n) represents the pixel grayscale value in the u-th row and n-th column of the image currently captured by the i-th node camera. The calculation requires traversing the entire image.

[0077] S23, Determine D i If (u,n) is greater than the threshold T, then I... i (u,n) is assigned the value 255, otherwise I is assigned the value 1. i (u,n) is assigned the value 0. The judgment and assignment need to traverse the entire image, where the threshold T is set according to the actual working scenario.

[0078] S24. Perform opening and dilation operations in the image morphology operations in sequence.

[0079] S25. Generate the smallest bounding box that encloses the connected components in the image, and output its center pixel coordinates (u). t ,n t ).

[0080] This invention provides the above-mentioned target detection algorithm based on the background model; other dynamic target detection algorithms can also be used here.

[0081] This invention provides the above-mentioned spatial positioning algorithm based on the two-dimensional coordinates of the target object in the imaging plane of the node camera. Other target positioning algorithms can also be used here.

[0082] Step five specifically includes the following steps:

[0083] S51. Calculate the deflection angle ψ of the projection of the intruding target onto the imaging plane of the camera at the i-th node. i1 and θ i1 , where ψ i1 θ is the horizontal deflection angle. i1 The vertical deflection angle:

[0084] ψ i1 =atan2(λ,(u t -u0)du),θ i1 =atan2(λ / cosψ i1 ,(n t -n0)dn);

[0085] Where λ represents the focal length of the node camera, (u0,n0) represents the center pixel coordinates of the acquired image, and du and dn represent the length and width of the pixels in the acquired image, respectively.

[0086] S52. Calculate the overall deflection angle ψ of the intruding target relative to the i-th node camera. i and θ i :

[0087] ψ i =ψi1 +ψ i2 ,θ i =θ i1 +θ i2 ;

[0088] Where ψ i2 and θ i2 These represent the horizontal and vertical deflection angles of the camera at the i-th node, respectively.

[0089] S53. Repeat the above steps to calculate the overall deflection angle ψ of the intruding target relative to the camera at the j-th node. j and θ j .

[0090] S54. Calculate the position coordinates (x, y) of the intruding target in three-dimensional space. ij ,y ij ,z ij ):

[0091] y ij =D / (tanψ) i -tanψ j ),x ij =y ij tanψ i ,z ij =y ij tanθ i / cosψ i ;

[0092] Where D represents the distance between the i-th node camera and the j-th node camera.

[0093] If N equals 2 in step four, then the actual position coordinates of the intruding target in three-dimensional space are equal to (x ij ,y ik ,z ik If N is greater than 2, the node cameras that detect the intrusion target in the acquired image are divided into pairs, and the position coordinates of the intrusion target in three-dimensional space are calculated according to the above algorithm. The actual position coordinates are taken as the average of the calculated position coordinates.

[0094] Step six specifically includes the following steps:

[0095] S61. The flight trajectory for intercepting the drone is designed as follows:

[0096]

[0097] Where a k Let T represent the coefficients of the k-th order polynomial, and T represent the time interval for flight trajectory planning.

[0098] S62. Determine the constraints, where the starting position Sd (0) represents the three-dimensional spatial coordinates of the intercepting drone, with the termination position S. d (T) represents the three-dimensional spatial position of the intrusion target. The velocity and acceleration constraints at the start and end times are determined according to the actual scenario.

[0099] S63. Based on the above constraints, solve the polynomial coefficients in the flight trajectory expression to determine the trajectory function.

[0100] This invention provides the above-mentioned trajectory planning algorithm based on a fifth-order polynomial, and other types of UAV flight trajectories can also be used here.

[0101] Step seven specifically includes the following steps:

[0102] S71, based on the expected flight trajectory S of the intercepting drone d Define the trajectory tracking error δ = SS d , where S is the actual location of the intercepted drone.

[0103] S72. The design of the trajectory tracking controller for intercepting unmanned aerial vehicles is as follows:

[0104] u=k p δ+k d dδ;

[0105] Where u represents the control input, k p and k d These represent adjustable control parameters.

[0106] The present invention provides the above-described proportional-derivative trajectory tracking controller, and other UAV trajectory tracking controllers can also be used here.

[0107] This invention aims to enable drones to intercept drones that intrude into no-fly zones by relying on a ground camera network. Without requiring its own onboard environmental perception module, the system calculates the state information of the intruding drone through the ground camera network, and uses this information, along with surrounding environmental information, to plan the trajectory of the intercepting drone. A controller is designed to efficiently track the flight trajectory, ultimately realizing an automated drone aerial interception system.

[0108] 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, and no reference numerals in the claims should be construed as limiting the scope of the claims.

[0109] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for autonomous interception of aerial drones based on a ground camera network, comprising the following steps: Step 1: Node cameras in the ground camera network monitor the no-fly zone and continuously collect images; Step 2: Using a target detection algorithm, detect intrusion targets in the images captured by each node camera and calculate the two-dimensional coordinates of the intrusion targets on the imaging plane of the node camera. Step 3: Record the number of node cameras that detected the same intrusion target. ; Step 4: Judgment Is it greater than or equal to 2? If so, proceed to step five; otherwise, return to step one. Step 5: Based on the distribution and orientation information of the node cameras in the ground camera network, calculate the position of the intrusion target in three-dimensional space using the two-dimensional coordinates of the intrusion target on the imaging plane of different node cameras in the ground camera network. Step Six: Based on the relative three-dimensional spatial positions of the intrusion target and the intercepting drone, design the interception trajectory for the intercepting drone; Step Six specifically includes: S61. Design an interception trajectory for intercepting drones. : ； express Polynomial coefficients of order 1 Indicates the time interval for flight trajectory planning; S62. Determine the constraints: including the starting position. To intercept the drone's three-dimensional spatial coordinates, the termination position The three-dimensional spatial position of the intrusion target, the velocity and acceleration constraints at the start time, and the velocity and acceleration constraints at the end time are determined according to the actual scenario; S63. Based on the constraints, solve for the polynomial coefficients in the interception trajectory to determine the interception trajectory; Step 7: Design an interceptor drone trajectory tracking controller to track the interception trajectory and control the interceptor drone to intercept the intruding target.

2. The method for autonomous interception of aerial unmanned aerial vehicles based on a ground camera network according to claim 1, characterized in that, Step two specifically includes the following steps: S21. In the absence of intrusive targets, establish background models for each node camera in the ground camera network, where the first... The background model built by the node camera is used It means that, with Indicates the first In the nth node camera background model OK, The pixel grayscale value of the column; S22. Each node camera continuously acquires images and calculates pixel grayscale differences. ,in Indicates the first The image currently acquired by the nth node camera is the th OK, The pixel grayscale value of the column; S23, Judgment Is it greater than the threshold? If so, then it is Assign a value of 255, otherwise, Assign the value 0; S24. Perform opening and dilation operations in the image morphology operations in sequence; S25. Generate the smallest bounding box that encloses the connected components in the image, and output the center pixel coordinates of the smallest bounding box. This means obtaining the two-dimensional coordinates of the intruding target on the imaging plane of the node camera.

3. The method for autonomous interception of aerial unmanned aerial vehicles based on a ground camera network according to claim 1, characterized in that, Step five specifically includes the following steps: S51, Calculate the first The horizontal deflection angle of the projection of the intruding target onto the imaging plane of each nodal camera. and the vertical deflection angle : ； in The function that calculates the azimuth angle. Indicates the focal length of the nodal camera. This represents the center pixel coordinates of the acquired image. and These represent the length and width of pixels in the captured image, respectively. The two-dimensional coordinates of the intrusion target on the imaging plane of the node camera; S52, Calculate the relative position of the intruding target to the first The overall deflection angle of the camera at each node and : ； in and They represent the first The angle of deflection of each node camera in the horizontal and vertical directions; S53. Repeat steps S51 and S52 to calculate the relative position of the intruding target to the first... The overall deflection angle of the camera at each node and ; S54. Calculate the position coordinates of the intruding target in three-dimensional space. : ； in Indicates the first The node camera and the first The distance between each node camera; S55, if in step four If the value equals 2, then the actual coordinates of the intrusion target in three-dimensional space are: ;like If the value is greater than 2, the node cameras that detect the intrusion target in the image are divided into pairs and grouped together. According to steps S51 to S54, the position coordinates of the intrusion target corresponding to each group of node cameras in three-dimensional space are calculated respectively. The actual position coordinates of the intrusion target in three-dimensional space are taken as the average of the position coordinates of the intrusion target corresponding to each group of node cameras in three-dimensional space.

4. The method for autonomous interception of aerial unmanned aerial vehicles based on a ground camera network according to claim 1, characterized in that: Step seven specifically includes: S71, based on the desired interception trajectory of the intercepting drone. Define trajectory tracking error ,in To intercept the actual location of the drone; S72, Design of a trajectory tracking controller for intercepting drones: ； in Indicates control input, and All of these indicate adjustable control parameters.

5. The method for autonomous interception of aerial unmanned aerial vehicles based on a ground camera network according to claim 1, characterized in that, The interception system used includes: The ground sensing unit, which is connected to the communication transmission unit, consists of cameras fixedly installed on the ground or buildings. Each camera acts as a node camera and is connected to other node cameras through data transmission lines, thus forming a ground camera network. Data from each node camera in the ground camera network is transmitted to the central computing unit through the communication transmission unit. The central computing unit, connected to the communication transmission unit, receives images captured by node cameras in the ground sensing unit, runs target detection algorithms to perceive intruding targets in the images captured by the node cameras, runs spatial positioning algorithms to calculate the three-dimensional spatial coordinates of the intruding target, runs trajectory planning algorithms to design the interception trajectory of the UAV based on the relative positional relationship between the intercepting UAV and the intruding target, runs tracking control algorithms to track the flight trajectory of the intercepting UAV, and finally, the control signals obtained by the tracking control algorithms are transmitted to the interception execution unit through the communication transmission unit. The communication transmission unit includes a wired transmission module and a wireless transmission module. The wired transmission module connects the ground sensing unit and the central computing unit through a data transmission line, transmitting the visual image data acquired by the ground sensing unit to the central computing unit for intrusion target detection. The wireless transmission module connects the central computing unit and the execution interception unit through radio signals, transmitting the trajectory tracking control signals generated by the central computing unit to the execution interception unit for interception flight. The interception unit, i.e. the interceptor drone entity, is connected to the communication transmission unit. Through the airborne flight controller, it converts the trajectory tracking control signal generated by the central computing unit into the rotation speed of the interceptor drone's actuators, thereby achieving the tracking and interception of the intruding target.

Images