Ground moving target detection method based on distributed floating airship platform

Abstract

The invention discloses a ground moving target detection method based on a distributed floating airship platform, and the method employs a distance time domain pulse coherence algorithm to achieve ground clutter simulation and moving target echo simulation based on the distributed floating airship platform. The echo signals received by the original N floating airship platforms are preprocessed; carrying out distributed multi-view ground moving target detection based on an RFT algorithm, and obtaining position information and speed parameter information of a detected target in a unified ground grid coordinate system; and by utilizing the obtained observation information data in the plurality of scanning periods and combining Kalman filtering and a nearest neighbor data association algorithm, realizing motion trail tracking prediction of the single target or the plurality of targets. According to the method, high-precision detection and parameter estimation of the ground moving target are realized by utilizing multi-view data fusion, and trajectory tracking of complex multiple targets is realized by matching with a target tracking algorithm based on Kalman filtering and nearest neighbor data association.

Description

Technical Field

[0001] The present invention belongs to the field of radar moving target detection, and in particular relates to a ground moving target detection method based on a distributed floating airship platform. Background Art

[0002] The integration of radar GMTI (Ground Moving Target Indication) technology with a multi-perspective distributed radar platform layout demonstrates the necessity and importance of modern radar technology in complex environments. The core of GMTI technology lies in its ability to monitor and identify ground moving targets in real time, which is crucial in modern military and security settings. As battlefield environments become increasingly complex, traditional single-perspective radar systems often struggle to fully and accurately capture the motion trajectories and behavior patterns of dynamic targets. Therefore, a multi-perspective distributed radar platform layout can effectively overcome this limitation. A multi-perspective distributed radar platform deploys multiple radar sensors at different locations, forming a collaborative network. This layout not only provides wider coverage but also enhances target identification and tracking capabilities by observing from different perspectives. Combined with GMTI technology, this layout can acquire target motion information in real time, improving the accuracy and reliability of target detection. In complex urban environments or changing battlefield conditions, distributed radar systems can provide more comprehensive situational awareness through multi-source data fusion, enabling commanders to make more informed decisions. Furthermore, with the widespread adoption of new platforms such as drones and satellites, the flexibility and adaptability of distributed radar systems have been greatly enhanced. By combining GMTI technology with a multi-view distributed radar platform, all-around monitoring of ground targets can be achieved, enabling timely detection of potential threats. This technological advancement not only enhances operational effectiveness but also promotes technological innovation in related fields, driving the development of technologies such as signal processing, data analysis, and artificial intelligence.

[0003] In summary, the combination of radar GMTI technology and a multi-perspective distributed radar platform deployment is both a necessary means of addressing modern security challenges and a crucial guarantee for enhancing national security and military competitiveness. With continued technological advancement, these two technologies will play an increasingly important role in future military and civilian applications. However, these technologies also face challenges: limited ground coverage for traditional single-platform GMTI systems, high R&D costs for distributed satellite-based GMTI systems, and limited sustained operational capabilities for distributed airborne GMTI systems. Summary of the Invention

[0004] Purpose of the invention: The present invention proposes a ground moving target detection method based on a distributed floating airship platform to fill the gap in the relevant multi-view target detection algorithm, and cooperates with the Kalman filter and nearest neighbor data association algorithm to ultimately achieve large-scale ground moving target detection, parameter estimation and trajectory tracking.

[0005] Technical solution: The present invention provides a method for detecting ground moving targets based on a distributed floating airship platform, comprising the following steps:

[0006] (1) Using the range-time-domain pulse coherence algorithm, the ground clutter simulation and the echo simulation of the moving target based on the distributed floating airship platform are realized, and Gaussian random noise is added;

[0007] (2) Preprocessing the echo signals received by the original N floating airship platforms;

[0008] (3) For the pre-processed echo signals obtained from all airship platforms, distributed multi-view ground moving target detection is performed based on the RFT algorithm, and the position information and velocity parameter information of the detected targets are obtained in a unified ground grid coordinate system;

[0009] (4) The motion trajectory tracking and prediction of a single target or multiple targets is achieved by utilizing the observation information data of multiple scanning cycles obtained through distributed multi-view ground motion target detection based on the RFT algorithm, combined with Kalman filtering and nearest neighbor data association algorithm.

[0010] Furthermore, the distributed floating airship platform is set in near space, 20-100 km above the ground.

[0011] Furthermore, the implementation process of step (2) is as follows:

[0012] After completing pulse compression in the range frequency domain, it is further transformed into the range Doppler domain. After the data at the Doppler center is set to 0, it is inversely transformed into the two-dimensional time domain to eliminate the interference of ground clutter.

[0013] Furthermore, the distributed multi-view ground moving target detection based on the RFT algorithm is implemented as follows:

[0014] The radar's time-domain transmitted signal is a linear frequency modulation signal p(t), and the target has a constant radial velocity relative to the radar platform. p(t) is expressed as:

[0015] p(t)=rect(t / T)exp(jπK r t 2 )

[0016] Among them, K ris the range modulation frequency, t is the time variable, T is the signal pulse width, and the echo signal received by the radar is expressed as follows after demodulation to baseband:

[0017]

[0018] Among them, τ and η are fast time and slow time respectively, A i is the amplitude of target i, R i (η)=R i0 +V i η, represents the relative distance between the i-th target and the radar platform at time η, V i is the radial velocity of target i, R i0 is the initial distance between target i and the radar platform, f c is the carrier frequency, and the range FFT of S1(τ,η) is performed to obtain:

[0019]

[0020] Among them, q(f) is the frequency domain form of p(t), S2(f,η) is further subjected to range pulse compression, and R i (η) is rewritten as R i0 +V i η is obtained:

[0021]

[0022] Convert the above formula to the time domain:

[0023]

[0024] At this time, the target envelope is affected by the parameter R i0 and V i Control, the purpose of RFT conversion is to detect the correct (V i ,R i0 ), in the known parameters (V i ,R i0 ) in the case of speed V i The corresponding frequency f i =2V i / λ is accumulated in the azimuth direction using Fourier transform; in the process, the distance to time is expressed as slant distance:

[0025] S5(R i0 ,V i )=∫S4[2(R i0 +V i η) / c,η]exp-j4πV i η / λ)dη

[0026] The energy accumulation value of the corresponding distance and radial velocity is obtained; since in actual situations the parameter pair (V i ,R i0 ) is unknown, and it is necessary to search for various possible parameters at specified intervals, and finally obtain a representation of the RFT transform based on the above derivation:

[0027] S6(R,V)=∫S4[2(R+Vη) / c,η]exp(-j4πVη / λ)dη

[0028] Obtain the combined energy value of each distance search point-radial distance search point on the RV two-dimensional plane, that is, obtain the two-dimensional search point energy matrix based on the RFT algorithm;

[0029] The search range and step size of distance and speed are pre-set. Among them, the distance search point directly adopts the one-way distance sampling point, and the speed search point is specifically divided into speed magnitude search point and speed direction search point. The speed direction search point is defined as the angle between the target's motion direction and a specified reference direction. The pre-processed signal under the perspective of each airship platform is processed separately by RFT processing to detect the initial position parameters and radial velocity parameters of the potential moving target in the echo information of each perspective.

[0030] Furthermore, the process of obtaining the position information and velocity parameter information of the detected target in the unified ground grid coordinate system is as follows:

[0031] The irradiated area of ​​the ground at the current wave position is divided into rectangular grids, and three nested loops are designed: the outermost layer traverses all grid points, the second layer traverses all view angles, and the innermost layer is a double loop that traverses all velocity magnitude and direction search points; the average one-way distance from the current grid point to the current view angle and the equivalent radial velocity of the current velocity search point at the current view angle are calculated, and these two parameters are used to match the corresponding energy accumulation value in the RFT two-dimensional search point energy matrix of the corresponding platform view angle; and it is temporarily stored in the result storage matrix corresponding to each view angle. When all view angles are traversed, the result storage matrices of all view angles are modulo-valued and superimposed on each other, and the peak value is taken as the final energy accumulation value of the grid point. At the same time, according to the position of the peak value in the matrix, the velocity magnitude and velocity direction corresponding to the energy value of the grid point are obtained. After storage is completed, the loop moves to the next grid point and the above loop operation is repeated;

[0032] After traversing all grid points, CFAR detection is performed on the final ground grid point energy matrix. By setting the decision threshold, the interference caused by noise and erroneous grid points is eliminated, and the initial position and two-dimensional velocity of the detected moving target are finally obtained.

[0033] A device according to the present invention includes a memory and a processor, wherein:

[0034] a memory for storing computer programs capable of running on the processor;

[0035] The processor is configured to execute the steps of the above-mentioned method for detecting ground moving targets based on a distributed floating airship platform when running the computer program.

[0036] The storage medium of the present invention stores a computer program, which, when executed by at least one processor, implements the steps of the above-mentioned ground moving target detection method based on a distributed floating airship platform.

[0037] Beneficial effects: Compared with the existing technology, the present invention has the following beneficial effects: 1. The present invention adopts a distributed floating airship platform as the carrier platform of the radar. Compared with a single platform architecture, the ground coverage of the present invention is wide and theoretically, as long as there are enough floating airship platforms, there is no detection blind spot; compared with a distributed satellite-borne platform, the research and development and operation and maintenance costs of the present invention are significantly lower, and it can achieve rapid network filling and deployment; compared with a distributed airborne platform, the strong endurance of the airship also ensures a longer continuous operation time, which can reach several months or even a year; in addition, through the stationary state of the platform relative to the ground, the main lobe clutter can be constrained to an extremely narrow Doppler frequency band, thereby fundamentally avoiding the problem of clutter diffusion of traditional moving platforms; 2. The present invention sets the distributed floating airship platform in near space, 20-100km above the ground, compared with the existing The distributed airborne platform has higher concealment and is not easily affected by the attack range of traditional surface-to-air defense facilities. In addition, the airflow disturbance in the near space is small, so the motion state stability of the platform is relatively strong. 3. The multi-perspective ground moving target detection algorithm proposed in the present invention is based on the RFT algorithm. A set of multi-perspective target detection algorithms suitable for distributed floating airship platforms is designed, breaking through the traditional idea of ​​matching first and then positioning, and proposing an integrated matching and positioning process. By utilizing multi-perspective data fusion, through the long-term coherent accumulation of the radar time domain echo signal after the range pulse compression, it can break through the limitation of Doppler resolution and realize high-precision detection and parameter estimation of ground moving targets, and cooperate with the target tracking algorithm based on Kalman filtering and nearest neighbor data association to realize the trajectory tracking of complex multiple targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is a geometric diagram of ground moving target detection based on four airship platforms;

[0039] Figure 2 The flowchart of the ground moving target detection and parameter estimation algorithm based on RFT under any number of floating airship viewing angles;

[0040] Figure 3 The following is the RFT processing result of the echoes received by the four airship platforms;

[0041] Figure 4 This is the ground grid point energy matrix diagram after the fusion of four perspective information;

[0042] Figure 5 This is the result diagram after CFAR detection of the ground grid point energy matrix;

[0043] Figure 6 It is the ground true velocity magnitude map corresponding to the detected ground moving targets;

[0044] Figure 7 It is the ground true velocity direction map corresponding to the detected ground moving target;

[0045] Figure 8 In order to utilize the observation data of a moving target with two intersecting trajectories within multiple azimuth scanning cycles, the trajectory tracking result diagram is obtained by using Kalman filtering combined with the nearest neighbor data association algorithm. DETAILED DESCRIPTION

[0046] The present invention will be further described in detail below with reference to the accompanying drawings.

[0047] The present invention proposes a ground moving target detection method based on a distributed floating airship platform. The specific implementation process is as follows:

[0048] In this embodiment, the number of viewing angles is specified to be 4, the beam scanning direction is 20.78°, i.e. the 114th wave position, and the remaining simulation parameters are shown in Table 1. The corresponding ground moving target detection geometric relationship is as follows: Figure 1 As shown, a one-transmit-four-receive transmission and reception method is adopted, in which the main transmitting airship is airship No. 2, which transmits radar signals, and all platforms receive echo signals together.

[0049] Table 1 Related simulation system parameters

[0050]

[0051]

[0052] Firstly, the range-time-domain pulse coherence algorithm is used to simulate ground clutter based on the distributed airship platform, and Gaussian random noise is added.

[0053] Secondly, suppose there are two moving targets within the ground illumination range of the 114th wave position. The information of the two moving targets is shown in Table 2, where the velocity direction is defined as the moving direction of the moving target and Figure 1 The angle between the positive direction of the X axis.

[0054] Table 2 Parameter information of two moving targets

[0055]

[0056] Then, after adding the information of the two moving targets to the echo information, the echo signals received by the original four floating airship platforms are preprocessed. After completing pulse compression in the range frequency domain, they are further transformed into the range Doppler domain. After setting the data at the Doppler center to 0, they are inversely transformed into the two-dimensional time domain to eliminate the interference of ground clutter.

[0057] Then, the pre-processed echo signals obtained from all floating airship platforms are processed as follows: Figure 2 The distributed multi-view ground moving target detection based on the RFT algorithm shown in FIG1 obtains the position information and velocity parameter information of the detected target in a unified ground grid coordinate system.

[0058] Assume that the radar's time-domain transmitted signal is a linear frequency modulation signal p(t), and the target has a constant radial velocity relative to the radar platform. For simplicity of analysis, an echo signal in one of the coherent processing unit times (CPI) is used as an example, where p(t) can be expressed as:

[0059] p(t)=rect(t / T)exp(jπK r t 2 )

[0060] Among them, K r is the range modulation frequency, t is the time variable, T is the signal pulse width, and the echo signal received by the radar can be expressed as follows after demodulation to baseband:

[0061]

[0062] Among them, τ and η are fast time and slow time respectively, A i is the magnitude of target i, where R i (η)=R i0 +V i η, represents the relative distance between the i-th target and the radar platform at time η, V i is the radial velocity of target i, R i0 is the initial distance between target i and the radar platform, f c is the carrier frequency, and performing range FFT on S1(τ,η) yields:

[0063]

[0064] Among them, q(f) is the frequency domain form of p(t), S2(f,η) is further subjected to range pulse compression, and R i (η) is rewritten as R i0 +Vi η is obtained:

[0065]

[0066] Convert the above formula to the time domain:

[0067]

[0068] At this time, the target envelope is affected by the parameter R i0 and V i Control, the purpose of RFT conversion is to detect the correct (V i ,R i0 ), in the known parameters (V i ,R i0 ) in the case of speed V i The corresponding frequency f i =2V i / λ is accumulated in the azimuth direction using Fourier transform. Note that the distance to time must be expressed as slant distance in the process, as shown in the following formula:

[0069] S5(R i0 ,V i )=∫S4[2(R i0 +V i η) / c,η]exp(-j4πV i η / λ)dη

[0070] So far, we have obtained the energy accumulation value of the corresponding distance and radial velocity, but in actual situations, the parameter (V i ,R i0 ) is unknown, so it is necessary to search for various possible parameters at specified intervals, and finally obtain a representation of the RFT transform based on the above derivation:

[0071] S6(R,V)=∫S4[2(R+Vη) / c,η]exp(-j4πVη)dη

[0072] Now, the combined energy value of each distance search point-radial distance search point on the RV two-dimensional plane is obtained, that is, the two-dimensional search point energy matrix obtained based on the RFT algorithm is obtained.

[0073] Assuming that there are N airship perspectives in total, the time domain echo signals obtained from each airship platform perspective after range pulse compression are subjected to separate RFT two-dimensional search processing, that is, the initial position parameters and radial velocity parameters of the potential moving target in the echo information of each perspective will be detected, and in this step, the search range and step size of the distance and speed need to be set in advance. Among them, the distance search point P directly uses the one-way distance sampling point, and the speed search point Q is specifically divided into the speed magnitude search point and the speed direction search point. The speed direction search point is defined as the angle between the target's moving direction and a specified reference direction. The final detection result is as follows: Figure 3 shown.

[0074] After obtaining the RFT two-dimensional search results corresponding to each view echo, the ground irradiated area at the current wave position is divided into M rectangular ground grids, and a three-layer nested loop is designed. First, the corresponding loop statistical variables are set. n represents the current view number, and its value traverses from 1 to the total number of view points N; m represents the current ground grid point number, and its value traverses from 1 to the total number of grid points M; q represents the current velocity search point number; S n The representative result storage matrix is ​​used to store the energy accumulation value extracted from the corresponding RFT processing result of each distance-speed two-dimensional search point in each perspective information.

[0075] The loop logic of the three-layer loop is as follows: the outermost loop traverses all grid points M, the second loop traverses all view angles N, and the innermost loop traverses all speed search points Q. However, since speed search points can be further divided into speed magnitude and speed direction search points, the innermost loop is actually a double loop that traverses all speed magnitude and direction search points. What the loop body actually does is to calculate the average one-way distance R from the current grid point m to the current view angle n after correction. m and the equivalent radial velocity V of the current velocity search point q at the current viewing angle m , and use these two parameters to match the corresponding energy accumulation value in the RFT two-dimensional search point energy matrix of the corresponding platform perspective, and temporarily store it in the result storage matrix S corresponding to each perspective n For the current perspective n, after traversing all the speed search points Q, enter the next perspective information to continue the loop operation. It should be noted that the current speed search point number q should be reset to 1 at this time, and Figure 2 The Y stands for "yes" and the N stands for "no".

[0076] For the current ground grid point m, after traversing all the viewing angles N, the result storage matrices S1 to S nTake the modulus values ​​and superimpose them, and take the peak value as the final energy accumulation value of the grid point. At the same time, according to the position of the peak value in the matrix, the speed magnitude and speed direction corresponding to the energy value of the grid point can be obtained. After storage, loop to the next grid point and repeat the above loop operation. It should be noted that the current perspective number n should be reset to 1 at this time.

[0077] After traversing all grid points M, the final ground grid point energy matrix can be obtained, such as Figure 4 As shown in the figure, CFAR detection is performed on it. By setting a reasonable decision threshold and eliminating the interference caused by noise and erroneous grid points, the initial position and two-dimensional velocity of the detected moving target can be obtained, as shown in the figure. Figure 5 、 Figure 6 and Figure 7 According to the detection results, the average positioning error is 1m×2m, the average velocity magnitude estimation error is 0.005m / s, and the average velocity direction estimation error is 0.5°.

[0078] The above is the process of detecting the status information of a ground moving target at a certain moment. If you want to track multiple targets, the specific implementation method is as follows:

[0079] The 66th wave position is selected as the observation angle, and the wave position sequence number can be flexibly adjusted according to the area to be observed. For the convenience of the embodiment, it is assumed that the ground moving target makes uniform linear motion in each azimuth scanning cycle, and 10 azimuth scanning cycle time periods are set. The speed and direction of the ground moving target are changed in each time period, and it is ensured that in these 10 azimuth scanning cycles, all moving targets are always within the illumination range of the current 66th wave position. Set a moving target with two intersecting trajectories, and its motion parameter information in 10 azimuth scanning cycles is as follows:

[0080] Table 3 Motion parameter settings of a ground moving target 1

[0081] Time period(s) Speed ​​(m / s) Speed ​​direction (°) End point coordinates (m) 0-1 11.2 153.4 (57648,90406) 1-2 11.2 206.5 (57638,90401) 2-3 11.2 243.4 (57633,90391) 3-4 11.2 243.4 (57628,90381) 4-5 11.2 296.5 (57633,90371) 5-6 11.2 26.5 (57643,90376) 6-7 11.2 26.5 (57653,90381) 7-8 10.0 270.0 (57653,90371) 8-9 11.2 243.4 (57648,90361) 9-10 11.2 333.4 (57658,90356)

[0082] Table 4 Motion parameter settings of a ground moving target 2

[0083] Time period(s) Speed ​​(m / s) Speed ​​direction (°) End point coordinates (m) 0-1 5 90 (57643,90351) 1-2 5 90 (57643,90356) 2-3 5 90 (57643,90361) 3-4 5 90 (57643,90366) 4-5 5 90 (57643,90371) 5-6 8 90 (57643,90379) 6-7 10 90 (57643,90389) 7-8 5 90 (57643,90394) 8-9 10 90 (57643,90404) 9-10 5 90 (57643,90409)

[0084] from Figure 8 As can be seen, the starting coordinates of Target 1 and Target 2 are separated by a certain distance. The overall trajectory diagram shows that the trajectories of Target 1 and Target 2 intersect at two points. However, confusion in the observed data only occurs when the trajectories intersect within the same or relatively close time periods. At the end of the sixth time period, the positions of the two targets are relatively close, which can easily cause confusion in the observed data.

[0085] According to the principle of Kalman filtering and nearest neighbor data association algorithm, Figure 2 The observation data of the two targets in Table 3 and Table 4 obtained by the algorithm flow in the above is used as input, and finally the following can be obtained: Figure 8 The results of tracking two moving targets are shown in Figure 1, where the red and blue solid lines represent the true motion trajectories of Target 1 and Target 2, respectively, while the purple and green dashed lines represent the detected motion trajectories of Target 1 and Target 2, respectively. It can be clearly seen that the trajectories of both moving targets are detected with high accuracy. In particular, at the intersection of the two true trajectories, thanks to the algorithm's high-precision parameter estimation of the two targets, the nearest neighbor data association algorithm successfully distinguishes the two targets at that moment, belonging to different motion trajectories.

[0086] The present invention also provides a device, including a memory and a processor, wherein: the memory is used to store a computer program that can be run on the processor; the processor is used to execute the steps of the ground moving target detection method based on the distributed floating airship platform as described above when running the computer program.

[0087] The present invention also provides a storage medium having a computer program stored thereon. When the computer program is executed by at least one processor, the steps of the ground moving target detection method based on the distributed floating airship platform are implemented.

[0088] The embodiments are only for illustrating the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. A ground moving target detection method based on a distributed floating airship platform, characterized in that: The following steps are involved: (1) Using the range-time-domain pulse coherence algorithm, the ground clutter simulation and the echo simulation of the moving target based on the distributed floating airship platform are realized, and Gaussian random noise is added; (2) Preprocessing the echo signals received by the original N floating airship platforms; (3) For the pre-processed echo signals obtained from all airship platforms, distributed multi-view ground moving target detection is performed based on the RFT algorithm, and the position information and velocity parameter information of the detected targets are obtained in a unified ground grid coordinate system; (4) The motion trajectory tracking and prediction of a single target or multiple targets is achieved by utilizing the observation information data of multiple scanning cycles obtained through distributed multi-view ground motion target detection based on the RFT algorithm, combined with Kalman filtering and nearest neighbor data association algorithm.

2. The method for detecting ground moving targets based on a distributed floating airship platform according to claim 1, characterized in that: The distributed floating airship platform is set in near space, 20-100 km above the ground.

3. The method for detecting ground moving targets based on a distributed floating airship platform according to claim 1, characterized in that: The implementation process of step (2) is as follows: After completing pulse compression in the range frequency domain, it is further transformed into the range Doppler domain. After the data at the Doppler center is set to 0, it is inversely transformed into the two-dimensional time domain to eliminate the interference of ground clutter.

4. The method for detecting ground moving targets based on a distributed floating airship platform according to claim 1, characterized in that: The implementation process of distributed multi-view ground moving target detection based on RFT algorithm is as follows: The radar's time-domain transmitted signal is a linear frequency modulation signal p(t), and the target has a constant radial velocity relative to the radar platform. p(t) is expressed as: p(t)=rect(t / T)exp(jπK r t 2 ) Among them, K r is the range modulation frequency, t is the time variable, T is the signal pulse width, and the echo signal received by the radar is expressed as follows after demodulation to baseband: Among them, τ and η are fast time and slow time respectively, A i is the amplitude of target i, R i (η)=R i0 +V i η, represents the relative distance between the i-th target and the radar platform at time η, V i is the radial velocity of target i, R i0 is the initial distance between target i and the radar platform, f c is the carrier frequency, and the range FFT of S1(τ,η) is performed to obtain: Among them, q(f) is the frequency domain form of p(t), S2(f,η) is further subjected to range pulse compression, and R i (η) is rewritten as R i0 +V i η is obtained: Convert the above formula to the time domain: At this time, the target envelope is affected by the parameter R i0 and V i Control, the purpose of RFT conversion is to detect the correct (V i ,R i0 ), in the known parameters (V i ,R i0 ) in the case of speed V i The corresponding frequency f i =2V i / λ is accumulated in the azimuth direction using Fourier transform; in the process, the distance to time is expressed as slant distance: S5(R i0 ,V i )=∫S4[2(R i0 +V i η)c,η]exp(-j4πV i the / the)dthe The energy accumulation value of the corresponding distance and radial velocity is obtained; since in actual situations the parameter pair (V i ,R i0 ) is unknown, and it is necessary to search for various possible parameters at specified intervals, and finally obtain a representation of the RFT transform based on the above derivation: S6(R,V)=∫S4[2(R+Vη) / c,η]exp(-j4πVη / λ)dη Obtain the combined energy value of each distance search point-radial distance search point on the RV two-dimensional plane, that is, obtain the two-dimensional search point energy matrix based on the RFT algorithm; The search range and step size of distance and speed are pre-set. Among them, the distance search point directly adopts the one-way distance sampling point, and the speed search point is specifically divided into speed magnitude search point and speed direction search point. The speed direction search point is defined as the angle between the target's motion direction and a specified reference direction. The pre-processed signal under the perspective of each airship platform is processed separately by RFT processing to detect the initial position parameters and radial velocity parameters of the potential moving target in the echo information of each perspective.

5. The method for detecting ground moving targets based on a distributed floating airship platform according to claim 1, characterized in that: The process of obtaining the position information and velocity parameter information of the detected target in the unified ground grid coordinate system is as follows: The illuminated area on the ground at the current wave position is divided into a rectangular grid, and a three-layer nested loop is designed: the outermost layer traverses all grid points, the second layer traverses all viewpoints, and the innermost layer is a double loop that traverses all velocity magnitude and direction search points. The average one-way distance from the current grid point to the current viewpoint and the equivalent radial velocity of the current velocity search point at the current viewpoint are calculated. These two parameters are used to match the corresponding energy accumulation value in the RFT two-dimensional search point energy matrix at the corresponding platform viewpoint. And temporarily store it in the result storage matrix corresponding to each view angle. When all view angles are traversed, the result storage matrices of all view angles are modulo and superimposed on each other, and the peak value is taken as the final energy accumulation value of the grid point. At the same time, according to the position of the peak value in the matrix, the speed magnitude and speed direction corresponding to the energy value of the grid point are obtained. After storage is completed, the cycle is cycled to the next grid point and the above cycle operation is repeated; After traversing all grid points, CFAR detection is performed on the final ground grid point energy matrix. By setting the decision threshold, the interference caused by noise and erroneous grid points is eliminated, and the initial position and two-dimensional velocity of the detected moving target are finally obtained.

6. A device, characterized in that: comprising a memory and a processor, wherein: a memory for storing computer programs capable of running on the processor; The processor is configured to execute the steps of the ground moving target detection method based on a distributed floating airship platform as claimed in any one of claims 1 to 5 when running the computer program.

7. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by at least one processor, implements the steps of the ground moving target detection method based on a distributed floating airship platform according to any one of claims 1 to 5.

Images