Space-borne SAR non-azimuth mode interferometric height measurement method

Abstract

The application discloses a spaceborne SAR non-ground-track mode interferometric height measurement method. The application establishes a re-track interferometric imaging processing model under a non-ground-track curved strip scene, designs a beam pointing range and a re-track interferometric baseline for a non-ground-track imaging mode configuration, optimally selects a non-ground-track local terrain re-track baseline design and a re-track image pair, and carries out interferometric phase processing and height measurement inversion on the re-track imaging image. The application proposes a design scheme for the problem of terrain height inversion by applying a satellite re-track interferometric capability under a spaceborne non-ground-track SAR imaging configuration, and makes up for the deficiency of the existing non-ground-track local terrain interferometric measurement capability.

Description

Technical Field

[0001] This invention relates to the field of synthetic aperture radar technology, specifically to a spaceborne SAR non-track mode interferometric elevation measurement method. Background Technology

[0002] Synthetic Aperture Radar (SAR) is an active microwave remote sensing device with advantages such as all-weather, all-time operation, high two-dimensional resolution, and strong penetration, making it significant for applications such as disaster early warning, environmental monitoring, and military reconnaissance. Non-track imaging mode is a unique operating mode of spaceborne SAR. Traditional spaceborne SAR imaging bands are generated along the satellite orbit, resulting in a single trajectory. Spaceborne non-track SAR, by continuously adjusting the elevation and azimuth beam pointing, directly generates an observation imaging band along the target terrain, rather than the traditional band generated along the satellite orbit. This fundamentally reduces echo data redundancy when imaging certain "non-track scenarios" such as seismic zones and coastlines, significantly improving the observation efficiency of spaceborne SAR for narrow scenes.

[0003] Interferometric Synthetic Aperture Radar (InSAR) is based on two-dimensional SAR data processing techniques. It combines two coherent complex images with satellite parameters and inverts the phase difference from the interferogram to calculate terrain information and deformation detection methods for the target area. This allows SAR imagery to be extended from two-dimensional to three-dimensional terrain scene information. Current InSAR processing is mostly based on spaceborne front-side-looking observation modes, and there is a lack of research on non-track-based observation modes. In the space-variable configuration of non-track imaging modes, due to the influence of slant geometry, orbital parameters, satellite attitude, and antenna beam variations, re-orbit observations using non-track modes may encounter problems such as non-parallel orbits and inconsistent slant angles. Conventional baseline selection methods based on equal Doppler centers cannot obtain good data correlation, and the InSAR data acquisition and processing methods of conventional spaceborne front-side-looking imaging modes cannot be directly applied.

[0004] Therefore, it is necessary to expand the application of interferometric SAR technology to non-track observation modes, study the difficulties of InSAR processing under this mode, and propose a spaceborne SAR non-track mode interferometric processing method to realize elevation inversion in non-track regions. Summary of the Invention

[0005] In view of this, the present invention provides a spaceborne SAR non-track mode interferometric elevation measurement method, which designs the beam pointing range and double orbit interferometric baseline for the non-track imaging mode configuration, and preferably acquires double orbit image pairs that constitute the interferometric processing capability, thereby realizing the planning of double orbit interferometric processing scheme for the imaging zone of curved scene in spaceborne non-track mode.

[0006] The spaceborne SAR non-track mode interferometric elevation measurement method of the present invention includes the following steps:

[0007] Step 1: Plan and design the satellite-borne non-track SAR slant position to minimize the slant range change of the SAR during the illumination of the target scene;

[0008] Step 2, select the optimal interference pair configuration:

[0009] The satellite uses double orbit observation for non-track target scenes. The beam pointing range of the double orbit observation is controlled according to the beam pointing range of the first orbit terrain observation, so that the double orbit observation range is consistent with the terrain observation range of the first orbit, ensuring that the Doppler center frequency difference between the two orbit images is minimized, and obtaining the optimal interferometric pair configuration for optimal azimuth Doppler decoherence.

[0010] Step 3: For the double-track observation data of the same non-track imaging area, the phase-preserving backward projection algorithm in the non-track mode is used to perform imaging processing to obtain SAR imaging images of the two tracks.

[0011] Step 4: Pre-filter the two-track SAR imaging images, filter out non-common spectra, perform image geometric coarse registration, pixel and sub-pixel level registration, generate interferograms and remove flat phase;

[0012] Step 5: After performing phase filtering on the interferogram after removing the flat phase, perform interferometric phase unwrapping processing;

[0013] Step 6: Using non-track satellite orbital track data and baseline parameters, perform elevation inversion and geocoding on the unwrapped interferometric phase to generate a non-track scene digital elevation model.

[0014] Preferably, in step 1, the target scene is first determined; then, based on the relative position between the spaceborne SAR and the target scene, a spatial geometric model is established, and the relationship between the geometric configuration parameters and the slant range variation range is obtained analytically; wherein, the slant range history of the target is represented by a Taylor series expansion slant range model; then, under the condition of satisfying the observation range of the target scene, the optimal beam pointing corresponding to the minimum slant range variation during the illumination of the scene under test is obtained by using a convex optimization method.

[0015] Preferably, in step 2, the critical baseline criterion for reorbit observation is:

[0016]

[0017] Where B is the system bandwidth, B ⊥ B is the vertical baseline. ⊥crit Let λ be the critical vertical baseline, c be the speed of light, and R be the slant range. Let θ be the beam angle, η be the beam down angle, and η be the terrain slope; the corresponding baseline decoherence criterion is:

[0018]

[0019] Based on the coherence of the heavy orbit interferometry capability and the beam pointing range in the satellite's non-track mode, control the decoherence of the heavy orbit spatial baseline and the azimuth Doppler decoherence;

[0020] The decoherence calculation method for the heavy rail space baseline is shown in Equation (6), where ρ r For range resolution:

[0021]

[0022] The method for calculating azimuth Doppler decoherence is shown in equation (7), ρ a For azimuth resolution, dφ is the difference in azimuth angle rotation between the two passing beams pointing towards the target scene:

[0023]

[0024] The total coherence coefficient is:

[0025] γ total =γ B ·γ D (8)

[0026] According to the coherence criterion shown in Equation (8), the double track interference baseline of the secondary overpass is selected. When designing the non-track double track beam, the azimuth beam angle geometry parameters are selected to control the consistency of the Doppler center frequency of the two beam illumination ranges, so that the azimuth Doppler decoherence reaches the best.

[0027] Preferably, the beam geometry parameters are selected within the range of 30 to 50° under beam-controlled viewing angle to control the effective baseline of the heavy track within 10 to 20% of the critical baseline constraint, and the spatial baseline decoherence parameters are controlled to ensure that the average coherence coefficient in the non-track observation interval is above 0.7.

[0028] Preferably, in step 3 of the imaging process, firstly, an imaging grid covering the entire scene is established along the extension direction of the curved scene observation zone according to the imaging resolution requirements; then, a grid that meets the beam coverage width is selected and offset according to the wave foot direction, with the offset amount in units of cells; the method of dividing the offset grid can reduce redundant imaging grids and improve the temporal imaging efficiency of non-track imaging; after the imaging grid is divided, the imaging grid coordinates are transformed from the scene coordinate system to the satellite antenna coordinate system, and the imaging grid is determined to be illuminated based on the beam illumination range results, and then the echo is back-projected onto the imaging grid to achieve coherent accumulation imaging.

[0029] Preferably, during the pre-filtering step 4, the spectral offsets in the range and azimuth directions under repeated orbit observations are first calculated:

[0030]

[0031] Where, Δf r , Δf a These are the range spectral offset and the azimuth spectral shift, respectively; f d1 and f d2 η represents the azimuth center frequency of the two-track imaging images, respectively. r v represents the slope of the terrain in the direction of distance; g Let η be the wave foot velocity. a Orientation to terrain;

[0032] Then, through spectrum truncation, non-common spectrum is filtered out, and the common interval of the spectrum is retained.

[0033] Preferably, in step 4, the coarse registration is as follows: First, the orbital trajectory of the main satellite is obtained by interpolation using the ephemeris data, and control points in the main image are selected and their geographic coordinates are obtained through forward geocoding; then, the orbital trajectory of the satellite in the secondary image is obtained through the same fitting method, and the image coordinates of the control points in the secondary image are obtained through backward geocoding; then, the image coordinates in the two images are subtracted to obtain the offsets in the azimuth and range directions, and the offsets between points in the main and secondary images are compensated based on the offsets.

[0034] The pixel-level registration is as follows: a target region is selected in the main image, coherence coefficients are calculated from the image to determine corresponding points, and polynomial fitting is performed based on the determined corresponding points to obtain the offset distribution of the entire image, thereby achieving pixel-level registration.

[0035] The subpixel-level registration is as follows: after completing pixel-level registration, non-baseband bilinear interpolation is first performed on the master and slave images, and then the coherence coefficient is calculated using a sliding window, and the offset fitting is completed to achieve subpixel-level registration.

[0036] Preferably, in step 5, the Goldstein filtering method is used for phase filtering, and the Goldstein splitting method is used for phase unwrapping.

[0037] A better approach is to untangle the phase as follows:

[0038] 1) Identify residual points in the interferogram. The specific identification method is as follows:

[0039]

[0040]

[0041] Where p and q are the row and column values ​​of the coordinates; For the entanglement operator, φ p,q Let d be the phase value of the pixel in row p and column q, when d ≠ When the value is 0, the point is considered a residual point, and the distribution of residual points in the interferogram is obtained accordingly.

[0042] 2) Scan the entire graph, find the first residual point, find the next residual point in the local area, and connect them. Repeat this process until the total number of positive and negative residual points in the connection is the same. Then end the connection and find a new starting residual point until all residual points are connected.

[0043] 3) Finally, integrate each pixel one by one around the branch tangent to untangle the image, avoiding the residual points being surrounded by the integration path, and finally obtain the untangled phase image.

[0044] Beneficial effects:

[0045] This invention establishes a double-track interferometric imaging processing model for non-track curved strip scenarios. The beam pointing range and double-track interferometric baseline are designed for the non-track imaging mode configuration. The optimal non-track local terrain double-track baseline design and double-track image pairs are selected. Interferometric phase processing and elevation measurement inversion are performed on the double-track imaging images. This invention proposes a design scheme for terrain elevation inversion using satellite double-track interferometric capabilities in spaceborne non-track SAR imaging configurations, overcoming the shortcomings of existing non-track local terrain interferometric measurement capabilities. Attached Figure Description

[0046] Figure 1 A flowchart of a spaceborne SAR non-track interferometry method provided by the present invention;

[0047] Figure 2 This is a schematic diagram of the re-orbit space configuration for spaceborne SAR non-track interferometry.

[0048] Figure 3 The influence of non-track beam angle parameters of spaceborne SAR on the critical baseline of heavy orbit.

[0049] Figure 4 The influence of non-track beam angle parameters of spaceborne SAR on the interferometric coherence coefficient. Detailed Implementation

[0050] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0051] This invention provides a method for spaceborne SAR non-track mode interferometric elevation measurement, the flowchart of which is shown below. Figure 1 As shown, it includes the following steps:

[0052] S1: Determine the non-track imaging zone region and plan and design the spaceborne non-track SAR wave positions;

[0053] Spaceborne non-track SAR heavy orbit configuration such as Figure 2 As shown, for specific scenarios, the satellite's observation imaging zone is no longer parallel to the nadir trajectory, and the scene exhibits certain bending and extension characteristics as well as an oblique tilt angle. In non-track imaging mode, it is necessary to continuously adjust the elevation and azimuth beam pointing to ensure that the observation imaging zone follows the target terrain direction.

[0054] The beam rotates in both elevation and azimuth dimensions. Assuming that at imaging time t... a The angle between the beam center and the satellite velocity is α(t). a When operating in non-track stripe mode, the angle between the beam center and the satellite velocity remains constant, i.e.:

[0055]

[0056] Meanwhile, based on the non-track SAR spatial configuration, the target's slant range history R(t) a The slant range model can be represented by a Taylor series expansion:

[0057] R(t a )=R0+k1(t a -t p )+k2(t a -t p ) 2 +k3(t a -t p ) 3 +k4(t a -t p ) 4 (2)

[0058] In the formula, t p Let t be the time at which the target synthetic aperture is centered. p The distance between the radar and the target at any given time, k1 to k4 are the coefficients of each order in the Taylor series expansion slant range model. Based on the relative position between the spaceborne SAR and the target scene, a spatial geometric model is established, and the relationship between the geometric configuration parameters and the slant range variation range is obtained analytically. Under the condition of satisfying the observation range of the "slant scene", the optimal beam pointing is obtained by using a convex optimization method to minimize the slant range variation during the illumination scene and reduce the time-varying degree of the slant range.

[0059] Spaceborne SAR non-track interferometry heavy orbit space configuration such as Figure 2 As shown, where β f β n These represent the maximum and minimum downward angle values ​​during the observation process, respectively; H is the orbital altitude; Ω is the angle between the nadir point trajectory and the target terrain direction, referred to as the scene oblique angle; W a W r These are the length and width of the diagonal scene, R.c α is the slant range value at the center time. α is the angle between the tangent to the ground and the scene azimuth direction formed by the plane formed by the antenna along the range direction, called the non-track observation angle.

[0060] Based on the criterion of minimizing slant range variation, and while satisfying the observation range of the oblique scene, the optimal non-track observation angle α′ is obtained using a convex optimization method, where W r0 For the desired distance-to-width ratio, W a0 R is the desired azimuth width. max R min These represent the maximum and minimum values ​​of the slope range change during the acquisition of the observed echo data, respectively, and their specific expressions are as follows:

[0061]

[0062] The optimal non-track observation angle α′ is obtained from equation (3). This allows us to determine the non-track observation angle α′ that minimizes slant range variation during data acquisition and reduces the time-varying effect of the slant range history under any orbit. S2: Select the optimal interferometric pair configuration based on the satellite reorbit baseline coherence criterion and beam pointing design;

[0063] After determining the non-track observation angle and slant range history, the beam pointing needs to be obtained based on the actual satellite antenna pointing in the satellite orbital auxiliary data, and the wave foot is calculated in conjunction with the orbital information from the two flights. After determining the beam pointing change and imaging range of the first orbit, the extension direction of the observation zone in the curved scene is determined according to the wave foot direction, and a ground reference point is selected to control the beam pointing change range of the double orbit observation. This ensures the consistency between the terrain observation range of the double orbit and the terrain observation range of the first orbit, and minimizes the difference in the Doppler center frequency of the two orbit observation data. This ensures that the overlapping area covered by the wave positions of the two non-track observations meets the interferometric coherence requirements.

[0064] In the oblique-view configuration of non-track imaging, the critical baseline criterion for re-orbit observation is stated as follows:

[0065]

[0066] Where B is the system bandwidth, B ⊥ B is the vertical baseline. ⊥crit Let λ be the critical vertical baseline, c be the speed of light, and R be the slant range. Let θ be the beam angle, η be the beam down angle, and η be the terrain slope. The corresponding baseline decoherence criterion is expressed in the following form:

[0067]

[0068] Based on the coherence of the heavy orbit interferometry capability and the beam pointing range in the satellite's non-track mode, control the decoherence of the heavy orbit spatial baseline and the azimuth Doppler decoherence.

[0069] The decoherence coefficient of the heavy rail spatial baseline is calculated as shown in equation (6), where B ⊥ For the vertical baseline, ρ r For range resolution:

[0070]

[0071] The azimuth Doppler decoherence coefficient is calculated as shown in equation (7), where θ is the downward angle and ρ is the vertical angle. a For azimuth resolution, dφ is the difference in azimuth angle rotation between the two passing beams pointing towards the target scene:

[0072]

[0073] Based on the above criterion for the correlation coefficient of multiple orbit interferometry, the total coherence coefficient is expressed as follows:

[0074] γ total =γ B ·γ D (8)

[0075] Where γ B γ D These represent the decorrelation caused by the double orbit spatial baseline and the two azimuth Doppler variations, respectively, and both are related to the changes in the beam angles of the master and slave tracks. According to Equation (6), the coherence criterion shown in Equation (7) is to select the double orbit interference baseline of the double orbit passing track. In the design of the non-track double orbit beam, the beam down angle and beam azimuth oblique angle parameters are selected to control the consistency of the Doppler center frequency of the two beam observation range images, so that the azimuth Doppler decorrelation coefficient reaches the optimum.

[0076] After adopting the optimal heavy-orbit beam configuration, the azimuth Doppler correlation coefficient γ D To achieve the optimal correlation coefficient, the correlation coefficient of the interferometric image pair depends primarily on the decoherence γ of the heavy orbit spatial baseline. B .

[0077] from Figure 3 It can be seen that when the beam azimuth angle constraint increases within the range of 10–30° in the non-track case, the critical baseline length decreases accordingly. When the beam down-angle constraint increases within the range of 30–50°, the critical baseline length increases accordingly. By selecting beam geometry parameters within the range of beam down-angle and beam azimuth angle constraints to control the effective baseline of the heavy track within 10–20% of the critical baseline constraint, and controlling the spatial baseline decoherence parameters to ensure that the average coherence coefficient in the non-track observation range is above 0.7, the overall map coherence quality and the ground elevation inversion accuracy of the interferometric processing are guaranteed.

[0078] S3: Select the double-track observation data of the same non-track imaging region and apply the non-track time-domain back projection algorithm for imaging processing;

[0079] To address the spatial variation and two-dimensional coupling problem in spaceborne SAR non-track imaging mode observations, a non-track imaging time-domain imaging algorithm can be adopted. This method has strong adaptability to both spatial and time-varying observation parameters.

[0080] The non-track imaging process employs a time-domain-based phase-preserving back-projection (BP) imaging algorithm. The BP algorithm can achieve SAR imaging under any spatial configuration, with no special requirements on the radar platform's orbital track or observation geometry. Furthermore, the BP algorithm exhibits good phase preservation, making it a reference algorithm for interferometry in non-track imaging mode. First, based on the imaging resolution requirements, an imaging grid covering the entire scene is established along the extension direction of the curved scene observation zone after two flyby observations. Then, a grid that meets the beam coverage width is selected and offset according to the wave foot direction, with the offset amount measured in cells. This method of dividing the offset grid reduces redundant imaging grids and improves the time-domain imaging efficiency of non-track imaging.

[0081] To address the issues of drastic Doppler center changes and spectral aliasing in non-track imaging due to two-dimensional beam scanning, the non-track imaging temporal domain imaging algorithm employs a refined beam determination method. This method transforms the imaging grid coordinates from the scene coordinate system to the satellite antenna coordinate system, determining whether the imaging grid is illuminated based on the beam illumination range. The echo is then back-projected onto the imaging grid to achieve coherent accumulation imaging. Wherein, T... syn To achieve a long synthesis time for the aperture, s(t,τ|x,y) represents the accumulated pixel signal, ultimately yielding the reconstructed results for each pixel in the imaging grid.

[0082]

[0083] In the back projection imaging process, the Doppler phase of the target is removed before coherent superposition. After coherent superposition, the Doppler phase needs to be recovered based on the slant distance of the scene observation zone center. Phase preservation processing is performed after coherent focusing in the azimuth direction of BP imaging. The slant distance from the aperture center to the target reference point is selected to perform phase compensation for each pixel value signal. The phase compensation term for each pixel value is as follows:

[0084]

[0085] S4: Perform pre-filtering on the two-track imaging images, complete image geometric coarse registration, pixel and sub-pixel level registration, generate interferogram and remove flat phase;

[0086] Based on the squint spatial variation configuration characteristics and imaging parameters of the non-track imaging mode, the basic steps for generating the interferometric phase from the re-orbit map of the non-track imaging image are as follows:

[0087] (1) Pre-filtering of heavy rail images

[0088] To correct the spectral shift caused by the difference in viewing angle, baseline, and Doppler parameters when two InSAR observations of the same target occur in non-track imaging, a pre-filtering process is performed before generating the interferometric phase of the re-track image.

[0089] The estimated spectral offsets in the range and azimuth directions under double orbit observations are as follows:

[0090]

[0091]

[0092] Among them, v g f is the wave foot velocity. d1 and f d2 These are the azimuth center frequencies of the two images, respectively. θ is the beam angle, θ is the beam down angle, and η is the beam angle. r η represents the slope of the terrain in the distance direction. a The azimuth slope is the topographic slope. The spectral offset can be directly calculated from the system parameters. Through spectral truncation, non-common spectra are filtered out, and common spectral regions are retained to ensure the coherence of subsequent interferometric images.

[0093] (2) Geometric registration of master and auxiliary images

[0094] This invention performs geometric coarse registration on two non-track imaging master and slave images, and uses the correlation function method to perform multi-level spatial registration at the pixel and sub-pixel levels.

[0095] 1. Coarse registration

[0096] First, the satellite's orbital trajectory is obtained by interpolation using the ephemeris data of the primary satellite. Control points in the primary image are then selected and geocoded forward to obtain their geographic coordinates. Next, the satellite's orbital trajectory in the secondary image is obtained using the same fitting method. The control points are then geocoded backward to obtain their image coordinates in the secondary image. Finally, subtracting the image coordinates from the two images yields the azimuth and range offsets, which are used to compensate for the offset between points in the primary and secondary images.

[0097] 2. Pixel-level registration

[0098] Pixel-level registration is performed using the coherence coefficient method, employing a sliding window approach. A target region is selected in the main image, and coherence coefficients are calculated from the image to determine corresponding points. Based on these determined corresponding points, a polynomial fitting is performed to obtain the offset distribution across the entire image, thus achieving pixel-level registration. The formula for calculating the coherence coefficient is as follows:

[0099]

[0100] Where Im1 and Im2 represent two SAR images, M and N represent the sliding window sizes, and u and v represent the sliding positions of the window. In practical applications, new coherence coefficients can also be constructed to contrast-stretch points with high coherence, thereby expanding the dynamic range.

[0101] 3. Subpixel-level registration

[0102] Because the Doppler center frequency of each target point deviates significantly from the zero-frequency position in the non-track imaging mode, non-baseband interpolation is used instead of baseband interpolation in the subpixel-level registration process. After completing pixel-level registration, non-baseband bilinear interpolation is first performed on the master and slave images, and then the coherence coefficient is calculated using a sliding window, and the offset fitting is completed.

[0103] (3) Generation of double orbit interferogram

[0104] After geometric registration, the master and slave images are processed to generate an interferometric phase map, followed by flat-ground effect removal. The interferometric phase expression generated by conjugate multiplication is as follows:

[0105]

[0106] Where Im1 and Im2 represent two SAR images, u int For the generated interferometric phase map, R is the slant range of the master image, and ΔR is the difference between the slant ranges of the master and slave images. The phase change after interferometry reflects the difference in slant ranges between the master and slave images.

[0107] The acquired orbital track data and heavy orbit baseline parameters are used for terrain clearing. A subset of control points are selected in the main image, and their actual coordinates and radar positions in the master and slave images are calculated. This yields the slant range difference between the master and slave images, and the phase difference of the control points is calculated. The phase difference expression is as follows:

[0108]

[0109] Finally, the flat-ground phase of the entire image is obtained by polynomial fitting using the phase difference of the control points, and then multiplied by the interferogram to remove the flat-ground phase.

[0110] S5: Perform phase filtering and unwrapping on the interferogram;

[0111] Common methods for phase filtering of interferograms after removing the flat-ground effect include the Goldstein branching method and the least squares method. In this case, the more robust Goldstein branching method is selected for filtering, which can reduce the number of phase residual points while maintaining the compactness of the interference fringes.

[0112] After phase filtering, interferometric phase unwrapping is performed using the Goldstein splitting method. The specific steps of the algorithm are as follows:

[0113] 1. Identify residual points in the interferogram. The specific identification method is as follows:

[0114]

[0115]

[0116] in For the entanglement operator, φ p,q Let d be the phase value of the pixel in row p and column q. When d ≠ 0, the point is considered a residual point. The distribution of residual points in the interferogram can be obtained in this way.

[0117] 2. Scan the entire graph to find the first residual point. Find the next residual point within a small range and connect them. Repeat this process until the total number of positive and negative residual points in the connection is the same. Then end the connection and find a new starting residual point until all residual points are connected.

[0118] 3. Finally, the unwrapped phase image is obtained by integrating each pixel one by one around the branch tangent to avoid the residual points being surrounded by the integration path.

[0119] S6: Based on system and baseline parameters, perform elevation inversion and geocoding to generate a digital elevation model for non-track scenes;

[0120] The unwrapped interferometric phase data was subjected to elevation inversion and geocoding using simulated non-track satellite orbital track data and baseline parameters. Elevation inversion employed a phase-to-elevation conversion method based on fuzzy elevation. Geocoding utilized a geolocation method based on a reference ellipsoid.

[0121] The phase-to-elevation conversion factor is the elevation ambiguity, and its calculation formula is as follows:

[0122]

[0123] Where R is the slant distance. θ is the beam angle, and B is the beam downward angle. ⊥ The vertical baseline. am This refers to the elevation ambiguity, which represents the elevation change corresponding to every 2π change in phase, h. refFor reference terrain elevation, the elevation information obtained from phase-to-elevation inversion is located in the radar image coordinate system (a, r, h). Geocoding is required to transform the elevation information from the image coordinate system to the geographic coordinate system. Below. Using the spatial coordinate system (x, y, z) as the basis for transformation, the transformed coordinates are as follows:

[0124] The final conversion yields the interferometric elevation processing results for the non-track observation area.

[0125] Therefore, this invention provides a method for a spaceborne SAR non-track interferometric measurement model. This invention solves the system design problem of the double-track interferometric configuration in spaceborne SAR non-track mode, and expands the ability to apply double-track InSAR technology for terrain elevation measurement in existing non-track interferometric imaging modes.

[0126] The simulation parameters of the critical baseline for heavy orbit interferometry of spaceborne non-track SAR are shown in Table 1.

[0127] Table 1 Simulation parameters of critical baseline for satellite-borne non-tracking SAR heavy orbit interferometry

[0128]

[0129] In this embodiment, the critical interferometric baseline and coherence coefficient criteria based on non-track imaging geometry parameters are used for evaluation. Figure 3 The influence of non-track beam angle parameters of spaceborne SAR on the critical baseline of heavy orbit is presented. Figure 4 The influence of non-track beam angle parameters on the interferometric coherence coefficient of spaceborne SAR is presented. Combined with... Figure 3 The variation trend of the critical baseline of the heavy track and the interferometric coherence in 4 shows that the non-track SAR heavy track interferometric elevation measurement task needs to combine the critical baseline constraint of the heavy track and the interferometric coherence criterion to guide the selection of the variation range of the non-track observation beam angle.

[0130] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A spaceborne SAR non-track mode interferometric elevation measurement method, characterized in that, Includes the following steps: Step 1: Plan and design the slant range of the spaceborne non-track SAR to minimize the slant range variation during the illumination of the target scene. Specifically, first, determine the target scene; then, based on the relative position between the spaceborne SAR and the target scene, establish a spatial geometric model and analyze the relationship between the geometric configuration parameters and the slant range variation range; the slant range history of the target is represented by a Taylor series expansion slant range model; then, under the condition of satisfying the observation range of the target scene, use a convex optimization method to obtain the optimal beam pointing corresponding to the minimum slant range variation during the illumination of the scene under test. Step 2, select the optimal interference pair configuration: The satellite uses double orbit observation for non-track target scenes. The beam pointing range of the double orbit observation is controlled according to the beam pointing range of the first orbit terrain observation, so that the double orbit observation range is consistent with the terrain observation range of the first orbit, ensuring that the Doppler center frequency difference between the two orbit images is minimized, and obtaining the optimal interferometric pair configuration for optimal azimuth Doppler decoherence. Specifically, the critical baseline criterion for reorbiting observations is as follows: Where B is the system bandwidth, B ⊥ B is the vertical baseline. ⊥crit Let λ be the critical vertical baseline, c be the speed of light, and R be the slant range. θ is the beam angle, η is the beam downward angle, and η is the terrain slope. The corresponding baseline decoherence criterion is: Based on the coherence of the heavy orbit interferometry capability and the beam pointing range in the satellite's non-track mode, control the decoherence of the heavy orbit spatial baseline and the azimuth Doppler decoherence; The decoherence calculation method for the heavy rail space baseline is shown in Equation (6), where ρ r For range resolution: The method for calculating azimuth Doppler decoherence is shown in equation (7), ρ a For azimuth resolution, dφ is the difference in azimuth angle rotation between the two passing beams pointing towards the target scene: The total coherence coefficient is: c total =c B ·c D (8) According to the coherence criterion shown in Equation (8), the heavy track interference baseline of the secondary overpass is selected. When designing the non-track heavy track beam, the azimuth beam angle geometry parameter is selected to control the consistency of the Doppler center frequency of the two beam illumination ranges, so that the azimuth Doppler decoherence reaches the best. Step 3: For the double-track observation data of the same non-track imaging area, the phase-preserving backward projection algorithm in the non-track mode is used to perform imaging processing to obtain SAR imaging images of the two tracks. Step 4: Pre-filter the two-track SAR imaging images, filter out non-common spectra, perform image geometric coarse registration, pixel and sub-pixel level registration, generate interferograms and remove flat phase; Step 5: After performing phase filtering on the interferogram after removing the flat phase, perform interferometric phase unwrapping processing; Step 6: Using non-track satellite orbital track data and baseline parameters, perform elevation inversion and geocoding on the unwrapped interferometric phase to generate a non-track scene digital elevation model.

2. The spaceborne SAR non-track mode interferometric elevation measurement method as described in claim 1, characterized in that, Under beam-controlled viewing angle constraints within the range of 30–50°, beam geometry parameters are selected to control the effective baseline of the heavy track within 10–20% of the critical baseline constraint. Spatial baseline decoherence parameters are controlled to ensure that the average coherence coefficient in the non-track observation interval is above 0.

7.

3. The spaceborne SAR non-track mode interferometric elevation measurement method as described in claim 1, characterized in that, In step 3 of the imaging process, firstly, an imaging grid covering the entire scene is established along the extension direction of the curved scene observation zone according to the imaging resolution requirements; then, a grid that meets the beam coverage width is selected and offset according to the wave foot direction, with the offset amount in units of cells; after the imaging grid is divided, the imaging grid coordinates are transformed from the scene coordinate system to the satellite antenna coordinate system, and the imaging grid is determined to be illuminated based on the beam illumination range results. Then, the echo is back-projected onto the imaging grid to achieve coherent accumulation imaging.

4. The spaceborne SAR non-track mode interferometric elevation measurement method as described in claim 1, characterized in that, In step 4, during pre-filtering, the spectral offsets in the range and azimuth directions under repeated orbit observations are first calculated: Where, Δf r , Δf a These are the range spectral offset and the azimuth spectral shift, respectively; f d1 and f d2 η represents the azimuth center frequency of the two-track imaging images, respectively. r v represents the slope of the terrain in the direction of distance; g Let η be the wave foot velocity. a Orientation to terrain; Then, through spectrum truncation, non-common spectrum is filtered out, and the common interval of the spectrum is retained.

5. The spaceborne SAR non-track mode interferometric elevation measurement method as described in claim 1, characterized in that, In step 4, the coarse registration is as follows: First, the orbital trajectory of the main satellite is obtained by interpolation using the ephemeris data, and control points in the main image are selected and their geographic coordinates are obtained through forward geocoding; then, the orbital trajectory of the satellite in the secondary image is obtained through the same fitting method, and the image coordinates of the control points in the secondary image are obtained through backward geocoding; then, the image coordinates in the two images are subtracted to obtain the offsets in the azimuth and range directions, and the offsets between points in the main and secondary images are compensated based on the offsets. The pixel-level registration is as follows: a target region is selected in the main image, coherence coefficients are calculated from the image to determine corresponding points, and polynomial fitting is performed based on the determined corresponding points to obtain the offset distribution of the entire image, thereby achieving pixel-level registration. The subpixel-level registration is as follows: after completing pixel-level registration, non-baseband bilinear interpolation is first performed on the master and slave images, and then the coherence coefficient is calculated using a sliding window, and the offset fitting is completed to achieve subpixel-level registration.

6. The spaceborne SAR non-track mode interferometric elevation measurement method as described in claim 1, characterized in that, In step 5, the Goldstein filtering method is used for phase filtering, and the Goldstein splitting method is used for phase unwrapping.

7. The spaceborne SAR non-track mode interferometric elevation measurement method as described in claim 6, characterized in that, Phase untangling specifically involves: 1) Identify residual points in the interferogram. The specific identification method is as follows: Where p and q are the row and column values ​​of the coordinates; For the entanglement operator, φ p,q Let d be the phase value of the pixel in row p and column q. When d ≠ 0, the point is considered a residual point. The distribution of residual points in the interferogram can be obtained in this way. 2) Scan the entire graph, find the first residual point, find the next residual point in the local area and connect them. Repeat this process until the total number of positive and negative residual points in the connection is the same. Then end the connection and find a new starting residual point until all residual points are connected. 3) Finally, integrate each pixel one by one around the branch tangent to untangle the image, avoiding the residual points being surrounded by the integration path, and finally obtain the untangled phase image.

Images