Distributed unmanned aerial vehicle mixed multi-baseline height inversion method of InSAR

Abstract

The application discloses a distributed unmanned aerial vehicle-borne InSAR mixed multi-baseline height inversion method, which is applied to the field of radar technology and mainly solves the problem that the existing technology cannot adapt to the configuration characteristics of flexible and variable baselines in a new type of distributed unmanned aerial vehicle-borne InSAR system, and the problem that the existing phase unwrapping-height inversion baseline configuration is greatly limited and is difficult to adapt to the new distributed system to perform height inversion on complex terrains; the implementation process comprises the following steps: (1) performing BP imaging on echo data obtained by a main platform and a sub-platform of the distributed unmanned aerial vehicle; (2) performing phase preprocessing operation on a SAR complex image, wherein the phase preprocessing operation comprises interference phase generation and non-local mean filtering operation; (3) establishing an optimization model and estimating the gradient of an interference phase ambiguity distance and a direction; (4) solving real ambiguity; (5) recovering real unwrapping interference phase according to the ambiguity and the wrapped phase; and (6) realizing high-precision height inversion according to a phase-height relationship.

Description

Technical Field

[0001] This invention belongs to the field of radar technology, and specifically relates to a high-precision elevation inversion technology. Background Technology

[0002] Interferometric Synthetic Aperture Radar (InSAR) is a crucial method for acquiring high-precision digital elevation models (DEMs) of the ground. Through a series of data processing steps, including SAR imaging, image registration, removal of terrain distortion, phase filtering, phase unwrapping, and elevation inversion, this technology can acquire high-precision DEMs of ground targets around the clock and in all weather conditions, enabling large-scale, high-efficiency mapping of these targets.

[0003] Compared to traditional airborne array InSAR systems and spaceborne multi-baseline InSAR systems, distributed UAV-borne InSAR systems offer superior flexibility and maneuverability. The flexible baseline characteristic allows for real-time adjustments to the UAV radar payload's flight trajectory to achieve flexible system configurations, meeting the mapping needs of diverse terrains. Furthermore, in battlefield environments, the decentralized, distributed, multi-platform airborne InSAR system provides platform redundancy; even if a few airborne platforms are destroyed, the remaining platforms can still complete the current detection mission. Therefore, hybrid multi-baseline InSAR elevation inversion based on distributed UAV-borne platforms has broad application prospects.

[0004] In distributed UAV-based scenarios, baseline parameters such as horizontal tilt angle are inconsistent, and the parameters corresponding to each baseline are not identical. Therefore, the interferometric phase and baseline length are not simply proportional. Furthermore, in many steep mountainous terrains, the absolute value of the phase difference between adjacent pixels in SAR images is not less than π, meaning the continuous phase assumption is no longer applicable in steep mountainous terrain. Therefore, traditional InSAR system multi-baseline phase unwrapping and elevation inversion methods cannot adapt to flexible baseline configurations and complex, steep observation terrains, and cannot be directly applied to distributed UAV-based systems for detecting steep mountainous terrain. In the paper "Interferometric synthetic aperture radar terrain elevationmapping from multiple observations" (Proceedings of IEEE 6th Digital Signal Processing Workshop, pp. 33-36, 1994), a baseline length-based phase mixing method is used to expand elevation ambiguity, thereby alleviating the limitations of the continuous phase assumption. However, for some excessively steep terrains, the continuous phase assumption is still not satisfied. These residual points that violate the continuous phase assumption will have a global impact on phase unwrapping, thus reducing the accuracy of elevation inversion. In the paper "An Optimization of Weighted Multi-Baseline LS Unwrapping Algorithm Based on Quality Map" (2019 IEEE International Geoscience and Remote Sensing Symposium, pp. 1697-1700, 2019), a phase difference gradient map is used to compensate for residual regions, mitigating the impact of residual points on global phase unwrapping. However, compensating for residual points does not fundamentally solve the limitation of the continuous phase assumption and remains unsuitable for steeply changing mountainous scenarios.The paper "Robust Two-Dimensional Phase Unwrapping for Multibaseline SAR Interferograms: A Two-Stage Programming Approach" (IEEE Transactions on Geoscience and Remote Sensing, vol. 54, no. 9, pp. 5217-5225, 2016) combines the multi-baseline approach (Chinese Remainder Theorem, CRT) and the single-baseline approach (Minimum Norm Method) to unwrap the entangled phase in two steps. In the multi-baseline stage estimation, the continuous phase assumption is broken, theoretically enabling elevation inversion for arbitrarily steep terrain. However, this method imposes many restrictions on InSAR system configurations, requiring all baseline inclination angles to remain consistent, meaning all UAV carriers must always be on a straight line, neglecting the need for baseline configuration differences and flexibility. Therefore, none of the above methods can achieve high-precision elevation inversion for complex and steep mountainous terrain in distributed UAV-borne InSAR systems. Summary of the Invention

[0005] To address the aforementioned technical issues, this invention proposes a distributed unmanned aerial vehicle (UAV) InSAR hybrid multi-baseline elevation inversion method. This method employs an improved two-step phase unwrapping algorithm combined with the specific baseline configuration of the distributed hybrid baseline InSAR system to unwrap the interferometric phases acquired by the system, thereby completing the elevation inversion of ground targets.

[0006] The technical solution adopted in this invention is as follows: a distributed UAV-borne InSAR hybrid multi-baseline elevation inversion method, based on a distributed UAV-borne InSAR system comprising: a UAV-borne radar platform as the main platform, several other distributed UAV-borne radar platforms as secondary platforms, and a target located in the imaging area. The main platform is equipped with radar transceiver antennas and operates in a self-transmitting and self-receiving mode. The secondary platforms are equipped with receiving antennas to receive radar echo signals from ground scattering points. The inversion includes:

[0007] S1. Perform time-domain back-projection imaging on the echo data obtained from the main platform and the secondary platform respectively to obtain their respective SAR complex images;

[0008] S2. Pair the SAR complex image of the main platform with the SAR complex image of each sub-platform and perform conjugate multiplication to obtain the original interferometric phase map corresponding to several baselines.

[0009] S3. Perform nonlocal mean filtering on the original interferometric phase maps corresponding to the several baselines obtained in step S2;

[0010] S4. Based on the original interferometric phase diagrams corresponding to several baselines after processing in step S3, and combined with the baseline configuration of the distributed unmanned aerial vehicle system, the phase ambiguity gradient is solved based on the Chinese remainder theorem.

[0011] S5. Estimate the true phase ambiguity matrix by minimizing the difference between the phase ambiguity gradient obtained in step S4 and the true ambiguity gradient.

[0012] S6. Based on the true phase ambiguity matrix estimated in step S5, and combined with the obtained entangled phase matrix, the true phase is recovered.

[0013] S7. Based on the actual phase obtained in step S6, the elevation information of the ground target is retrieved.

[0014] The beneficial effects of this invention are as follows: This invention employs an improved two-step phase unwrapping algorithm combined with the specific baseline configuration of a distributed hybrid baseline InSAR system to unwrap the interferometric phases acquired by the system, thereby completing the elevation inversion of ground targets. The advantage of this invention is that, compared with existing phase unwrapping-elevation inversion methods, it considers the configuration characteristics of distributed UAV-borne InSAR systems, overcomes the limitations of baseline configuration in the original algorithm, and improves adaptability to new distributed UAV-borne systems. This invention can be applied to fields such as Earth remote sensing, resource exploration, geological mapping, and military reconnaissance. Attached Figure Description

[0015] Figure 1 This is a geometric structure diagram of the present invention.

[0016] Figure 2 This is a flowchart of the method provided by the present invention.

[0017] Figure 3 This is the specific configuration of the distributed unmanned aerial vehicle-borne InSAR system adopted in a specific embodiment of the present invention.

[0018] Figure 4 This is a simulated topographic map observed by the system in step one of the specific embodiments of the present invention.

[0019] Figure 5 This invention is used in a distributed unmanned aerial vehicle (UAV) InSAR system to generate a DEM by inverting the elevation of a simulated topographic map. Detailed Implementation

[0020] To facilitate understanding of the technical content of this invention by those skilled in the art, the following description, in conjunction with the accompanying drawings, further illustrates the invention.

[0021] This invention provides a distributed unmanned aerial vehicle (UAV) InSAR hybrid multi-baseline elevation inversion method, such as... Figure 2 As shown, the implementation process includes the following steps:

[0022] 1: Theoretical Analysis and Preparation

[0023] 11: Geometry of a Distributed Unmanned Aerial Vehicle-borne InSAR System

[0024] The flight altitude of the UAV-borne radar platform is denoted as H. The flight speed of the InSAR platform along the y-direction is denoted as V. The main platform is equipped with both radar transceiver antennas and operates in a self-transmitting and self-receiving mode. The Q secondary platforms are equipped with only receiving antennas to receive radar echo signals from ground scattering points. The radar incident angle is denoted as θ. The slant range from the main platform to the scene target is denoted as R0, and the slant range from the i-th secondary platform to the scene target is denoted as R. i The horizontal inclination angle of each baseline is denoted as α. i The elevation information of the ground target point P is denoted as h. A typical distributed UAV-borne hybrid multi-baseline InSAR elevation inversion system is as follows: Figure 1 As shown.

[0025] The value of Q is greater than or equal to 2.

[0026] A baseline is defined as the connection between a primary platform and a secondary platform.

[0027] 12: Phase Relationship of Distributed Unmanned Aerial Vehicle-borne InSAR Systems

[0028] In a distributed UAV-borne hybrid baseline InSAR system, the echo delay phase of the ground target point received by the main platform's receiving antenna can be expressed as:

[0029]

[0030] Where λ represents the wavelength, and R0 represents the slant range from the main platform radar antenna to the ground target point P. The scattering phase is represented by the scattering phase, which is considered an invariant in distributed unmanned aerial vehicle (UAV) InSAR systems.

[0031] The echo delay phase received by the i-th sub-platform receiving antenna from the ground target point P can be expressed as:

[0032]

[0033] Among them, record Let R be a vector consisting of the echo delay phases of Q sub-platforms to a point target P. P =[R1,R2,…,R Q ], R P This represents a vector consisting of the slant ranges from the Q sub-platforms to the point target P. By subtracting the phase components of the SAR complex image acquired by the main platform from the phase components of the SAR complex images acquired by each sub-platform, the absolute interferometric phase of a point P on the ground target can be obtained.

[0034]

[0035] Among them, R 0,P This represents the slant distance from the main platform to the target point P;

[0036] according to Figure 1 The spatial geometric relationships of the distributed multi-baseline UAV-borne InSAR system shown can be derived using the law of cosines as follows:

[0037]

[0038] B i Indicates the length of the i-th baseline;

[0039] By combining the SAR imaging projection principle with equations (3) and (4) for approximation, the phase height relationship of the distributed multi-baseline InSAR system can be obtained:

[0040]

[0041] Where h is the elevation information matrix of the ground target scene, ψ i Let ψ be the interferometric phase matrix corresponding to the i-th baseline in a distributed multi-baseline system. i The phase ψ of a point target P It is derived by expanding it into a surface target.

[0042] 2: Phase Preprocessing for Distributed Unmanned Aerial Vehicle-borne InSAR System

[0043] 21: Distributed Platform Radar Echo BP Imaging

[0044] The echo data obtained from Q+1 distributed UAV-borne platforms are subjected to time-domain back-projection imaging (BP imaging) to obtain Q+1 complex SAR images S0, S1, S2, S3, S4, S5, S6, S7, S8, S9, S1, S1, S1, S1, S2, S1, S1, S2, S3 ...1, S1, S2, S1 i (i = 1, 2, ..., Q);

[0045]

[0046] Where i = 0 represents the main platform information, i = 1, 2, ..., Q represents the secondary platform information, and A i The envelope of the SAR image. The phase information of a SAR image can be represented as:

[0047]

[0048] Where λ represents wavelength, R i This represents the slant range from the radar antenna of the i-th platform to the ground target. This represents the initial scattering phase and can be considered an invariant.

[0049] 22: Generation of interferometric phase between primary and secondary platforms

[0050] The SAR complex image S0 of the main platform obtained after BP imaging is compared with the SAR complex image S0 of the i-th sub-platform. i By performing conjugate multiplication on each pair of baselines, the original interferometric phase map corresponding to the i-th baseline in the system can be obtained.

[0051] 23: Nonlocal Mean Interference Phase Filtering (NLM)

[0052] The interference phase map corresponding to the i-th baseline is passed through a nonlocal mean filter to obtain the denoised and wrapped phase φ. i This process filters out interference phase noise while maintaining good edges to ensure the accuracy of subsequent steps such as phase unwrapping.

[0053] 3: Hybrid multi-baseline phase unwrapping of joint system baseline configuration

[0054] The hybrid multi-baseline phase unwrapping method consists of two steps: In the first step, the gradient of the phase ambiguity number k is calculated based on the Chinese remainder theorem, taking into account the baseline configuration of the distributed unmanned aerial vehicle (UAV) system; in the second step, the value of the phase ambiguity number is estimated based on a global optimization algorithm, combined with the single-baseline approach, thereby extracting the unwrapped phase. The detailed steps are summarized below:

[0055] 31: Baseline configuration of the joint system and solving the phase ambiguity gradient using the Chinese remainder theorem

[0056] For the same ground target, the system inversion elevation should be consistent, therefore equation (5) can be rewritten as:

[0057]

[0058] Where φ i The entangled interferometric phase corresponding to the i-th baseline can be directly obtained through the InSAR system; k i Let represent the fuzzy number matrix corresponding to the i-th baseline.

[0059] For the pixels (m-1,n), (m,n-1), and (m,n) in the image pixel matrix, equation (8) can be expressed as:

[0060]

[0061] In the equation (9) above the distance, we have:

[0062]

[0063] in This represents the phase gradient in the range direction. This represents the fuzzy number gradient in the distance direction.

[0064] By minimizing the deviation between both sides of equation (10), the solution of the equation can be transformed into an optimization problem:

[0065]

[0066] Here, int represents an integer.

[0067] The azimuth direction is handled in the same way as the distance direction.

[0068] In the first step, combining the distributed UAV-borne baseline configuration, a multi-baseline architecture is used to solve for the phase ambiguity gradients in the range and azimuth directions:

[0069] 32: Recovery of fuzzy information based on global optimization

[0070] Based on the blur number gradient at pixel (m,n) obtained in step S21 Minimize the gradient of the true fuzzy number and The difference between them is used to estimate the true phase ambiguity matrix:

[0071]

[0072] Where f(·) is the specific objective function, These are the weighting coefficients for the distance and azimuth directions, respectively. In this embodiment, the weighting coefficient is set to 1. f(·) is selected as the L-1 norm (minimum cost flow) to ensure that the optimal solution is an integer, thus eliminating the integer restriction in equation (12).

[0073] 33: Deriving the untangling phase

[0074] The relationship between the entanglement phase and the true unentanglement phase can be expressed as:

[0075] ψ=φ+2π×k (13)

[0076] By combining the entangled phase matrix φ obtained by the system with the phase ambiguity matrix k obtained in steps 31 and 32, the true phase can be recovered to retrieve the elevation information of the ground target.

[0077] 4: Elevation inversion to generate a digital elevation model (DEM)

[0078] The unwrapped phase obtained by the multi-baseline phase unwrapping in step 3 is combined with Equation (5) for phase height conversion processing, and the unwrapped interference phase is converted into elevation information to recover the high-precision digital elevation model.

[0079] This embodiment primarily employs simulation experiments for verification, and all steps and conclusions have been verified correctly using Matlab 2022. The following provides a further detailed description of the specific implementation methods.

[0080] Step 1: Establish the spatial geometry of the distributed UAV-borne hybrid multi-baseline InSAR. Based on the geometry and the target in the imaging area, generate the simulated echo matrices of the ground targets from the main platform and each sub-platform of the distributed UAV-borne SAR system, denoted as S0(t,τ), S... i (t,τ), i=1,2,…,Q, the parameters required for simulation are shown in Table 1. The specific system configuration required for simulation is as follows. Figure 3 As shown, the target scenario is as follows Figure 4 As shown.

[0081] Table 1 System Parameters

[0082] Baseline configuration / m [0,1.3,2.4,3.2,4.6] Carrier height / m 2000 Beam tilt angle / ° 30 Baseline tilt angle / ° 15 Aircraft speed / m / s 20 Radar movement speed / m / s 20

[0083] like Figure 3 As shown in Table 1, a total of 5 drones are shown. The baselines from the drone serving as the main platform to the other 4 drones serving as secondary platforms are d1=1.3, d2=2.4, d3=3.2, and d4=4.6, respectively.

[0084] Step 2: Process the echo matrix S generated in Step 1 i BP imaging is performed on (t,τ), i=0,1,2,…,Q, and the SAR complex image matrices of the main and secondary platforms are denoted as S0 and S1, respectively. i ,i=1,2,…,Q.

[0085] Step 3: Since the BP algorithm integrates registration, terrain effect removal, and imaging, the registration and terrain effect removal operations can be omitted. Therefore, the BP-SAR complex image S0 obtained from the main platform in Step 2 and the BP-SAR complex images S0 obtained from each sub-platform are compared separately. i Paired conjugate multiplication is performed on i = 1, 2, ..., Q to obtain the initial entangled interference phase map corresponding to the i-th baseline.

[0086] Step 4: The interferometric phase maps obtained in Step 3 are filtered through a nonlocal mean filter (NLM) with a neighborhood window of 7×7, a search window of 15×15, a Gaussian smoothing factor of 5, and a Gaussian weighted distance Gaussian kernel of 0.5 to obtain the interferometric phase matrix φ after removing phase noise. i .

[0087] Step 5: Establish range equations based on the Chinese Remainder Theorem and baseline configuration parameters of a distributed UAV-borne hybrid multi-baseline InSAR system. The range phase gradient Distance gradient of fuzzy numbers The equation-solving problem is transformed into an optimization problem by minimizing the deviation between both sides of the equation:

[0088]

[0089] By solving the above optimization problem, the fuzzy number gradient in the range direction can be estimated. The same method can be used to solve for the azimuth ambiguity gradient.

[0090] Step Six: Minimize the gradient of the true blur number at pixel (m,n) and the gradient of the blur number estimated in Step Five. The following optimization problem is proposed for solution:

[0091]

[0092] Where f(·) is the L-1 norm.

[0093] The phase ambiguity number matrix k can be obtained by solving the above optimization problem. r ,r=1,2,…,Q.

[0094] Step 7: Based on the relationship between the wrapped phase and the true phase ψ r =φ r +2π×k r The hybrid multi-baseline unwrapped phase ψ is recovered. r .

[0095] Step 8: By combining the relationship between elevation and the unwound true phase, the elevation information can be recovered.

[0096]

[0097] High-precision 3D and 2D DEMs obtained by inversion, such as Figure 5 As shown.

[0098] Table 2 shows the error of elevation inversion of simulated topographic maps using the present invention in a distributed UAV-borne InSAR system. It can be seen that the present invention can achieve high-precision elevation inversion of steep mountain terrain under the new distributed UAV-borne InSAR system.

[0099] Table 2 shows the error of the present invention in the elevation inversion of simulated topographic maps.

[0100]

[0101] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Various modifications and variations can be made to the invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the invention should be included within the scope of the claims of the invention.

Claims

1. A distributed unmanned aerial vehicle (UAV) InSAR hybrid multi-baseline elevation inversion method, characterized in that, The distributed unmanned aerial vehicle (UAV) InSAR system includes: a UAV-borne radar platform as the main platform, several other distributed UAV-borne radar platforms as secondary platforms, and a target located in the imaging area. The main platform is equipped with a radar transceiver antenna and operates in a self-transmitting and self-receiving mode. The secondary platforms are equipped with receiving antennas to receive radar echo signals from ground scattering points. The inversion method includes: S1. Construct the phase height relation for a distributed multi-baseline InSAR system; S2. Perform time-domain back-projection imaging on the echo data obtained from the main platform and the secondary platform respectively to obtain their respective SAR complex images; S3. Pair the SAR complex images of the main platform with the SAR complex images of each sub-platform and perform conjugate multiplication to obtain the original interferometric phase maps corresponding to several baselines. S4. Perform nonlocal mean filtering on the original interferometric phase maps corresponding to the several baselines obtained in step S3 to obtain the denoised wrapped phase. S5. Based on the several entangled phases processed in step S4, and combined with the baseline configuration of the distributed unmanned aerial vehicle system, the phase ambiguity gradient is solved based on the Chinese remainder theorem. S6. Estimate the true phase ambiguity matrix by minimizing the difference between the phase ambiguity gradient obtained in step S5 and the true ambiguity gradient. S7. Based on the true phase ambiguity matrix estimated in step S6, and combined with the entangled phase matrix, the true phase is recovered. S8. Based on the actual phase obtained in step S7 and the phase height relationship of the distributed multi-baseline InSAR system constructed in step S1, the elevation information of ground objects is inverted.

2. The distributed unmanned aerial vehicle (UAV) InSAR hybrid multi-baseline elevation inversion method according to claim 1, characterized in that, Step S1 specifically includes the following sub-steps: S11. The echo delay phase received by the main platform's receiving antenna from the ground target point is expressed as: ； in, Indicates wavelength. Indicates the distance from the main platform's radar antenna to the ground target point. The slope distance, Indicates the scattering phase; S12, No. Each secondary platform receiving antenna received data from ground target points. The echo delay phase is expressed as: ； in, , This indicates the total number of secondary platforms. , Indicates by A vector consisting of the echo delay phases of each sub-platform to a point target P, denoted as... , Indicates by The vector consisting of the slant distances from each sub-platform to the target point P. For the first The slant distance from each secondary platform to the scene target; S13. Subtract the phase portion of the SAR complex image obtained from the main platform and the phase portion of the SAR complex images obtained from each sub-platform after BP imaging to obtain the ground target points. Absolute interference phase: ； ； in, This represents the slant distance from the main platform to the target point P; S14. Based on the spatial geometric relationships of the distributed multi-baseline UAV-borne InSAR system, and combined with the cosine theorem, we obtain: ； ； in, Indicates the first The length of the baseline, H represents the radar incident angle, and H represents the flight altitude of the UAV-borne radar platform. S15. Combining the SAR imaging projection principle with the formulas obtained in steps S13 and S14, an approximation is performed to obtain the phase height relationship of the distributed multi-baseline InSAR system: ； in, This is a matrix of elevation information for the target ground features. For the first distributed multi-baseline system The interferometric phase matrix corresponding to each baseline, Depend on It is derived by expanding it into a surface target; ɑ i This indicates the horizontal tilt angle of the baseline.

3. The distributed unmanned aerial vehicle (UAV) InSAR hybrid multi-baseline elevation inversion method according to claim 2, characterized in that, Step S5 is as follows: S51. For the same ground target, the system inversion elevation is consistent, thus the formula obtained in step S15 can be rewritten as follows: ； ； ； in, Q represents the number of secondary platforms; Indicates the first The entanglement interference phase corresponding to the baseline, Indicates the first The fuzzy number matrix corresponding to each baseline; S52, Targeting pixels in the image pixel matrix , , The formula obtained in step S51 is expressed as: ； ； S53. Combining the formulas obtained in step S52 with the formulas in the distance direction, we have: ； in, This represents the phase gradient in the range direction. Represents the fuzzy gradient of the distance upwards; S54. By minimizing the deviation between the two sides of the equation in step S53, the solution of the equation is transformed into an optimization problem: ； ； in, Represents an integer; S55. The treatment of azimuth is consistent with that of distance. S56. Combining the distributed UAV-borne baseline configuration, the phase ambiguity number gradients in the range and azimuth directions are solved using a multi-baseline architecture: .

4. The distributed unmanned aerial vehicle (UAV) InSAR hybrid multi-baseline elevation inversion method according to claim 3, characterized in that, Step S6 is as follows: Based on the solution obtained in step S56 Blur gradient at pixel Minimize the gradient of the true fuzzy number and The difference between them is used to estimate the true phase ambiguity matrix: ； ； in, Let be the objective function. These are the weighting coefficients for distance and azimuth, respectively. Choose the L-1 norm.

5. A distributed unmanned aerial vehicle (UAV) InSAR hybrid multi-baseline elevation inversion method according to claim 4, characterized in that, The expression for step S7 is: ； in, This is the winding phase matrix obtained from several winding phases after processing in step S4. This represents the phase ambiguity number matrix obtained in step S6.

Images