Single unmanned aerial vehicle target tracking method and system based on improved box particle filter algorithm

Abstract

The invention discloses a single unmanned aerial vehicle target tracking method and system based on an improved box particle filter algorithm, and belongs to the field of unmanned aerial vehicle target tracking. According to the method, the box particle filter algorithm is improved, the improved box particle filter algorithm is low in time complexity and high in real-time performance, more accurate and more robust target tracking can be realized, and the problems of low precision and poor real-time performance of each traditional interval analysis method in the box particle filter updating process of the existing tracking algorithm under the condition of a single unmanned aerial vehicle are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle target tracking, and in particular to a single unmanned aerial vehicle target tracking method and system based on an improved box particle filter algorithm. Background Art

[0002] In the field of drone target tracking, unmanned aerial vehicles (UAVs) are widely used in aerial photography, agriculture, plant protection, express delivery, disaster relief, wildlife observation, infectious disease monitoring, surveying and mapping, news reporting, power inspection and military fields. With the increase in drone applications, the demand for drone target tracking technology is also growing, especially the ability to track targets based on non-visual sensor data.

[0003] UAV target tracking technology has been widely used in military reconnaissance, environmental monitoring, traffic monitoring and other fields. Existing target tracking algorithms, especially vision-based methods, often have difficulty maintaining stable and accurate tracking in complex environments (such as fast movement, illumination changes, occlusion, etc.). In addition, the computing resources and energy supply of UAV platforms are limited, which puts higher requirements on the real-time performance and computing efficiency of tracking algorithms.

[0004] Traditional particle filters (PF) are flexible in dealing with nonlinear and non-Gaussian problems, but the algorithm needs to generate a large number of random particles, which has poor real-time performance. Moreover, over time, particle filters will experience particle degradation, that is, the weights of most particles become insignificant, and only a few particles carry almost all the weights, resulting in a decrease in sample diversity and affecting tracking performance. Summary of the invention

[0005] The present invention provides a single UAV target tracking method and system based on an improved box particle filter algorithm. The box particle filter algorithm is improved. The improved box particle filter algorithm has low time complexity and high real-time performance, and can achieve more accurate and robust target tracking. The problem that in the existing tracking algorithm, in the update process of the box particle filter, various traditional interval analysis methods have low accuracy and poor real-time performance in the case of a single UAV is solved.

[0006] The first aspect of the present invention provides a single UAV target tracking method based on an improved box particle filter algorithm, comprising the following steps:

[0007] Obtain measurement data sent back by the drone’s non-visual sensors;

[0008] A relative polar coordinate system between the drone and the target is established with the drone as the origin, and the box particle information and state space boundary of the box particle filter algorithm are modified according to the non-visual measurement data, so that the modified box particle interval and the measurement interval are both part of the spherical shell of the sphere;

[0009] In the updating process of the box particle filter algorithm, when the box particles shrink, it is determined whether all or part of the eight vertices of each box particle interval are in the measurement interval, and the vertex coordinates in the measurement interval and the boundary of the measurement interval are combined to form a new spherical shell to obtain the shrunk box particles;

[0010] The volume of each box particle after shrinkage is calculated as the weight of the box particle, and the state is estimated after normalization. After reaching the tracking step number, the UAV non-visual target tracking result is obtained.

[0011] Optionally, in one embodiment of the present invention, the measurement data returned by the non-visual sensor includes: distance, azimuth and elevation angle.

[0012] Optionally, in one embodiment of the present invention, modifying the box particle information and state space boundary of the box particle filter algorithm according to the non-visual measurement data includes:

[0013] The box particle information is represented by [X,Y,Z,X length ,Y length ,Z length ]Improved to The state space boundary is given by [X min ,X max ; Y min ,Y max ; Z min ,Z max ]Improved to Among them, X, Y, and Z of the box particle information before improvement are the Cartesian coordinates of the box particle. length ,Y length ,Z length is the side length of each dimension of the box particle, and the improved box particle information R,θ, is the coordinate of the box particle in the relative polar coordinate system, R length ,θ length , is the length of each dimension of the box particle, and the state space boundary before improvement [X min ,X max ; Y min ,Y max ; Z min ,Z max ] are the minimum and maximum values ​​of each dimension in the Cartesian coordinate system, respectively. The improved state space boundary They are the minimum and maximum values ​​of distance, azimuth and elevation angle in the relative polar coordinate system, respectively.

[0014] The second aspect of the present invention provides a single UAV target tracking system based on an improved box particle filter algorithm, comprising:

[0015] The acquisition module is used to obtain the measurement data sent back by the non-visual sensor of the drone;

[0016] An improved module is used to establish a relative polar coordinate system between the drone and the target with the drone as the origin, and to modify the box particle information and state space boundary of the box particle filter algorithm according to the non-visual measurement data, so that the modified box particle interval and the measurement interval are both part of the spherical shell of the sphere;

[0017] The shrinking module is used to determine whether all or part of the eight vertices of each box particle interval are within the measurement interval when the box particles shrink during the updating process of the box particle filter algorithm, and to form a new spherical shell with the vertex coordinates within the measurement interval and the boundary of the measurement interval to obtain the shrunk box particles;

[0018] The tracking module is used to calculate the volume of each box particle after contraction as the weight of the box particle, perform state estimation after normalization, and obtain the drone non-visual target tracking result after reaching the tracking step number.

[0019] Optionally, in one embodiment of the present invention, the measurement data returned by the non-visual sensor includes: distance, azimuth and elevation angle.

[0020] Optionally, in one embodiment of the present invention, modifying the box particle information and state space boundary of the box particle filter algorithm according to the non-visual measurement data includes:

[0021] The box particle information is represented by [X,Y,Z,X length ,Y length ,Z length ]Improved to The state space boundary is given by [X min ,X max ; Y min ,Y max ; Z min ,Z max ]Improved to Among them, X, Y, and Z of the box particle information before improvement are the Cartesian coordinates of the box particle. length ,Y length ,Z length is the side length of each dimension of the box particle, and the improved box particle information R,θ, is the coordinate of the box particle in the relative polar coordinate system, R length ,θ length , is the length of each dimension of the box particle, and the state space boundary before improvement [X min ,X max ; Y min ,Y max ; Z min ,Z max] are the minimum and maximum values ​​of each dimension in the Cartesian coordinate system, respectively. The improved state space boundary They are the minimum and maximum values ​​of distance, azimuth and elevation angle in the relative polar coordinate system, respectively.

[0022] An embodiment of the third aspect of the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to execute the single UAV target tracking method based on the improved box particle filter algorithm as described in the above embodiment.

[0023] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to perform the single UAV target tracking method based on the improved box particle filter algorithm as described in the above embodiment.

[0024] The single UAV target tracking method and system based on the improved box particle filter algorithm of the embodiment of the present invention improves the box particle filter algorithm in the state space boundary and the box particle initialization process. The improved box particle intervals and measurement intervals are all "spherical shells" of a sphere. Therefore, during the update process, when the box particles shrink, it is only necessary to determine whether the eight vertices of each box particle interval are all or partially within the measurement interval, and then the vertex coordinates in the measurement interval and the boundary of the measurement interval form a new "spherical shell" to obtain the shrunk box particles. Since the obtained shrunk box particles are also spherical shells, the volumes of the shrunk box particles can be directly calculated by mathematical methods as their weights, and the state estimation can be directly performed after normalization, omitting the process of finding the likelihood function and using the likelihood function to find the weights of each box particle, simplifying the algorithm flow, and since the improved shrinking step only has interval addition and subtraction, the precision loss of the interval shrinking algorithm can be avoided. Therefore, the improved box particle filter algorithm has low time complexity and high real-time performance, and can achieve more accurate and robust target tracking.

[0025] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0027] Figure 1 A flowchart of a single UAV target tracking method based on an improved box particle filter algorithm provided according to an embodiment of the present invention;

[0028] Figure 2Schematic diagram of the state space and box particles before improvement of an embodiment of the present invention;

[0029] Figure 3 Schematic diagram of the improved state space and box particles of an embodiment of the present invention;

[0030] Figure 4 A schematic diagram of a measurement interval space according to an embodiment of the present invention;

[0031] Figure 5 This is the flow chart of the classic box particle filter algorithm;

[0032] Figure 6 It is a flow chart of the improved box particle filter algorithm of an embodiment of the present invention;

[0033] Figure 7 An example diagram of a single UAV target tracking system based on an improved box particle filter algorithm according to an embodiment of the present invention;

[0034] Figure 8 A schematic diagram of the structure of an electronic device provided by an embodiment of the invention. DETAILED DESCRIPTION

[0035] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0036] In the field of drone target tracking, traditional visual tracking methods rely on image information captured by cameras, but in complex environments, such as low illumination, bad weather, or target occlusion, visual information may be unreliable or unavailable. Therefore, non-visual sensors, such as electromagnetic wave sensors, laser sensors fused with IMU sensors, etc., are valued for their robustness in complex environments. These sensors can provide information such as the distance, azimuth, and pitch angle of the target, providing a new solution for drone target tracking.

[0037] Box Particle Filter (BPF) uses box particles instead of traditional point particles. Each box particle represents a rectangular interval in the state space, thereby reducing the number of particles required and the computational complexity. However, in the actual application of UAV target tracking, since the non-visual sensors on the drone often transmit the distance, azimuth and pitch angle information of the target at each moment, due to the existence of measurement noise, this information is often an interval rather than a specific value, thus forming a measurement interval of a "spherical shell" part of the sphere, and the box particle interval is a rectangular parallelepiped. Therefore, in the update shrinkage process, interval shrinkage methods such as Gaussian elimination and constrained propagation are generally used to obtain the intersection of the box particle and the measurement interval. Among them, Gaussian elimination has the problem of low precision, and constrained propagation will produce dimensionality disasters, thus affecting the real-time performance of the code.

[0038] Therefore, in the contraction process of the classic box particle filter, the mathematical logic is complex and the code is highly complex, resulting in the problem of relatively poor real-time performance of the algorithm.

[0039] The present invention proposes an improved box particle filter algorithm for non-visual target tracking of UAVs, which is used to process the measurement data such as distance, azimuth, pitch angle, etc. transmitted back by the non-visual sensors on the UAVs to achieve more accurate and robust target tracking. In theory, there will be no loss of accuracy during the contraction process and the real-time performance will be greatly improved.

[0040] Specifically, Figure 1 The present invention provides a flowchart of a single UAV target tracking method based on an improved box particle filter algorithm according to an embodiment of the present invention.

[0041] like Figure 1 As shown, the single UAV target tracking method based on the improved box particle filter algorithm includes the following steps:

[0042] In step S101 , measurement data transmitted by a non-visual sensor of a drone is obtained.

[0043] In an embodiment of the present invention, the measurement data returned by the non-visual sensor includes: distance, azimuth angle, and pitch angle, etc.

[0044] In step S102, a relative polar coordinate system between the drone and the target is established with the drone as the origin, and the box particle information and state space boundary of the box particle filter algorithm are modified according to the non-visual measurement data, so that the modified box particle interval and the measurement interval are both part of the spherical shell of the sphere.

[0045] The present invention improves the state space boundary and the box particle initialization process, establishes the relative polar coordinate system between the drone and the target with the drone as the origin, and the box particle information is [X, Y, Z, Xlength ,Y length ,Z length ]Improved to The state space boundary is given by [X min ,X max ; Y min ,Y max ; Z min ,Z max ]Improved to Among them, X, Y, and Z of the box particle information before improvement are the Cartesian coordinates of the box particle. length ,Y length ,Z length is the side length of each dimension of the box particle, and the improved box particle information R,θ, is the coordinate of the box particle in the relative polar coordinate system, R length ,θ length , is the length of each dimension of the box particle, and the state space boundary before improvement [X min ,X max ; Y min ,Y max ; Z min ,Z max ] are the minimum and maximum values ​​of each dimension in the Cartesian coordinate system, respectively. The improved state space boundary are the minimum and maximum values ​​of distance, azimuth and elevation angle in the relative polar coordinate system. The schematic diagram of the box particles and measurement interval in the state space before and after the improvement is as follows: Figure 2 and Figure 3 shown.

[0046] In step S103, during the update process of the box particle filter algorithm, when the box particles shrink, it is determined whether the eight vertices of each box particle interval are all or partially within the measurement interval, and the vertex coordinates within the measurement interval and the boundary of the measurement interval are combined to form a new spherical shell to obtain the shrunk box particles.

[0047] In step S104, the volume of each box particle after contraction is calculated as the weight of the box particle, and the state is estimated after normalization. After reaching the tracking step number, the drone non-visual target tracking result is obtained.

[0048] like Figure 4As shown in the figure, the improved box particle intervals and measurement intervals are all "spherical shells" of the sphere. Therefore, during the update process, when the box particles shrink, it is only necessary to determine whether all or part of the eight vertices of each box particle interval are in the measurement interval, and then the vertex coordinates in the measurement interval and the boundary of the measurement interval form a new "spherical shell" to obtain the shrunk box particles. Since the shrunk box particles obtained are also spherical shells, the volumes of the shrunk box particles can be directly calculated by mathematical methods as their weights, and the state estimation can be directly performed after normalization, omitting the process of finding the likelihood function and using the likelihood function to find the weights of each box particle, simplifying the algorithm flow. At the same time, since the improved shrinking step only has interval addition and subtraction, the accuracy loss of the interval shrinkage algorithm can be avoided. However, in the improved algorithm, during the prediction step, the box particles will have a slight deformation after the state transfer, resulting in the problem of inconsistent volumes before and after. However, since the volume of the box particles and the volume difference caused by the deformation are very different, this part of the new error is so small that it can be ignored compared with the error of the interval analysis algorithm.

[0049] The classical box particle filter algorithm and the algorithm flow of the present invention are as follows: Figure 5 and Figure 6 In the classic box particle filter algorithm, the prediction step is an iterative process, and the contraction step is also an iterative process, so the time complexity of the entire algorithm is o(n 3 ). In the improved box particle filter algorithm, there is only one iteration of the prediction step in the entire process, so the time complexity of the improved algorithm is o(n 2 ), which is an order of magnitude higher than the former.

[0050] Compared with the extended Kalman filter, the present invention has lower requirements on initial conditions on the basis of ensuring real-time performance, and since there is no process of linearizing the nonlinear system, there is no linearization error, so the tracking accuracy is higher.

[0051] Compared with the unscented Kalman filter, the present invention is less sensitive to noise and has low requirements on user experience because it does not need to introduce a self-set sigma point.

[0052] Compared with the classical box particle filter, the present invention has lower time complexity and higher real-time performance.

[0053] According to the single UAV target tracking method based on the improved box particle filter algorithm proposed in the embodiment of the present invention, in the shrinking process of the classic box particle filter, the mathematical logic is complex and the code complexity is high, resulting in relatively poor real-time performance of the algorithm. An improved box particle filter algorithm is proposed to reduce the algorithm time complexity from o(n 3 ) is reduced to o(n2 ), which greatly improves the real-time performance of the algorithm.

[0054] Next, a single UAV target tracking method system based on an improved box particle filter algorithm proposed according to an embodiment of the present invention is described with reference to the accompanying drawings.

[0055] Figure 7 This is an example diagram of a single UAV target tracking method system based on an improved box particle filter algorithm according to an embodiment of the present invention.

[0056] like Figure 7 As shown, the single UAV target tracking method system 10 based on the improved box particle filter algorithm includes: an acquisition module 100, an improvement module 200, a contraction module 300 and a tracking module 400.

[0057] Among them, the acquisition module 100 is used to obtain the measurement data sent back by the non-visual sensor of the drone. The improvement module 200 is used to establish a relative polar coordinate system between the drone and the target with the drone as the origin, and to modify the box particle information and state space boundary of the box particle filter algorithm according to the non-visual measurement data, so that the modified box particle interval and the measurement interval are both part of the spherical shell of the sphere. The contraction module 300 is used to determine whether the eight vertices of each box particle interval are all or partially within the measurement interval when the box particles are contracted during the update process of the box particle filter algorithm, and to form a new spherical shell with the vertex coordinates within the measurement interval and the boundary of the measurement interval to obtain the contracted box particles. The tracking module 400 is used to calculate the volume of each contracted box particle as the weight of the box particle, perform state estimation after normalization, and obtain the non-visual target tracking result of the drone after reaching the tracking step number.

[0058] Optionally, in one embodiment of the present invention, the measurement data returned by the non-visual sensor includes: distance, azimuth and elevation angle.

[0059] Optionally, in one embodiment of the present invention, modifying the box particle information and state space boundary of the box particle filter algorithm according to the non-visual measurement data includes:

[0060] The box particle information is represented by [X,Y,Z,X length ,Y length ,Z length ]Improved to The state space boundary is given by [X min ,X max ; Y min ,Y max ; Z min ,Z max ]Improved to Among them, X, Y, and Z of the box particle information before improvement are the Cartesian coordinates of the box particle.length ,Y length ,Z length is the side length of each dimension of the box particle, and the improved box particle information R,θ, is the coordinate of the box particle in the relative polar coordinate system, R length ,θ length , is the length of each dimension of the box particle, and the state space boundary before improvement [X min ,X max ; Y min ,Y max ; Z min ,Z max ] are the minimum and maximum values ​​of each dimension in the Cartesian coordinate system, respectively. The improved state space boundary They are the minimum and maximum values ​​of distance, azimuth and elevation angle in the relative polar coordinate system, respectively.

[0061] It should be noted that the aforementioned explanation of the embodiment of the single UAV target tracking method based on the improved box particle filter algorithm is also applicable to the single UAV target tracking system based on the improved box particle filter algorithm of this embodiment, and will not be repeated here.

[0062] Figure 8 A schematic diagram of the structure of an electronic device provided by an embodiment of the invention. The electronic device may include:

[0063] A memory 801 , a processor 802 , and a computer program stored in the memory 801 and executable on the processor 802 .

[0064] When the processor 802 executes the program, the single UAV target tracking method based on the improved box particle filter algorithm provided in the above embodiment is implemented.

[0065] Furthermore, the electronic device further comprises:

[0066] The communication interface 803 is used for communication between the memory 801 and the processor 802 .

[0067] The memory 801 is used to store computer programs that can be executed on the processor 802 .

[0068] The memory 801 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0069] If the memory 801, the processor 802 and the communication interface 803 are implemented independently, the communication interface 803, the memory 801 and the processor 802 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0070] Optionally, in a specific implementation, if the memory 801, the processor 802 and the communication interface 803 are integrated on a chip, the memory 801, the processor 802 and the communication interface 803 can communicate with each other through an internal interface.

[0071] The processor 802 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0072] This embodiment also provides a computer-readable storage medium on which a computer program is stored, characterized in that when the program is executed by a processor, the single UAV target tracking method based on the improved box particle filter algorithm is implemented as above.

[0073] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0074] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0075] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0076] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0077] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

Claims

1. A single UAV target tracking method based on an improved box particle filter algorithm, characterized in that: The following steps are involved: Obtain measurement data sent back by the drone’s non-visual sensors; A relative polar coordinate system between the drone and the target is established with the drone as the origin, and the box particle information and state space boundary of the box particle filter algorithm are modified according to the non-visual measurement data, so that the modified box particle interval and the measurement interval are both part of the spherical shell of the sphere; In the updating process of the box particle filter algorithm, when the box particles shrink, it is determined whether all or part of the eight vertices of each box particle interval are in the measurement interval, and the vertex coordinates in the measurement interval and the boundary of the measurement interval are combined to form a new spherical shell to obtain the shrunk box particles; The volume of each box particle after shrinkage is calculated as the weight of the box particle, and the state is estimated after normalization. After reaching the tracking step number, the UAV non-visual target tracking result is obtained.

2. The method according to claim 1, characterized in that The measurements returned by non-visual sensors include: distance, azimuth and elevation.

3. The method according to claim 1, characterized in that Modifications to the box particle information and state space boundaries of the box particle filter algorithm based on non-visual measurement data include: The box particle information is represented by [X,Y,Z,X length ,Y length ,Z length ]Improved to The state space boundary is given by [X min ,X max ; Y min ,Y max ; Z min ,Z max ]Improved to Among them, X, Y, and Z of the box particle information before improvement are the Cartesian coordinates of the box particle. length ,Y length ,Z length The length of each dimension of the box particle, the improved box particle information is the coordinate of the box particle in the relative polar coordinate system, is the length of each dimension of the box particle, and the state space boundary before improvement [X min ,X max ; Y min ,Y max ; Z min ,Z max ] are the minimum and maximum values ​​of each dimension in the Cartesian coordinate system, respectively. The improved state space boundary They are the minimum and maximum values ​​of distance, azimuth and elevation angle in the relative polar coordinate system, respectively.

4. A single UAV target tracking system based on an improved box particle filter algorithm, characterized in that: include: The acquisition module is used to obtain the measurement data sent back by the non-visual sensor of the drone; An improved module is used to establish a relative polar coordinate system between the drone and the target with the drone as the origin, and to modify the box particle information and state space boundary of the box particle filter algorithm according to the non-visual measurement data, so that the modified box particle interval and the measurement interval are both part of the spherical shell of the sphere; The shrinking module is used to determine whether all or part of the eight vertices of each box particle interval are within the measurement interval when the box particles shrink during the updating process of the box particle filter algorithm, and to form a new spherical shell with the vertex coordinates within the measurement interval and the boundary of the measurement interval to obtain the shrunk box particles; The tracking module is used to calculate the volume of each box particle after contraction as the weight of the box particle, perform state estimation after normalization, and obtain the drone non-visual target tracking result after reaching the tracking step number.

5. The system according to claim 4, characterized in that The measurements returned by non-visual sensors include: distance, azimuth and elevation.

6. The system according to claim 4, characterized in that Modifications to the box particle information and state space boundaries of the box particle filter algorithm based on non-visual measurement data include: The box particle information is represented by [X,Y,Z,X length ,Y length ,Z length ]Improved to The state space boundary is given by [X min ,X max ; Y min ,Y max ; Z min ,Z max ]Improved to Among them, X, Y, and Z of the box particle information before improvement are the Cartesian coordinates of the box particle. length ,Y length ,Z length The length of each dimension of the box particle, the improved box particle information is the coordinate of the box particle in the relative polar coordinate system, is the length of each dimension of the box particle, and the state space boundary before improvement [X min ,X max ; Y min ,Y max ; Z min ,Z max ] are the minimum and maximum values ​​of each dimension in the Cartesian coordinate system, respectively. The improved state space boundary They are the minimum and maximum values ​​of distance, azimuth and elevation angle in the relative polar coordinate system, respectively.

7. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the single UAV target tracking method based on the improved box particle filter algorithm as described in any one of claims 1 to 3.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement a single UAV target tracking method based on an improved box particle filter algorithm as described in any one of claims 1 to 3.

Images