Spaceborne SAR non-track mode differential interferometric deformation measurement method

Abstract

This invention discloses a differential interferometric deformation measurement method for spaceborne SAR non-track mode. The invention establishes a double-track interferometric imaging processing model for non-track scenarios, enabling interferometric phase processing and terrain measurement inversion of local terrain. This invention primarily addresses the problem of terrain deformation measurement using the double-track interferometric capability based on the three-track measurement method in spaceborne non-track SAR configurations, proposing a design scheme to fill the gap in local terrain deformation measurement capabilities in non-track modes.

Description

Technical Field

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

[0002] Non-track imaging mode is a unique operating mode of spaceborne SAR (Synthetic Aperture Radar). Compared to traditional spaceborne SAR imaging modes, non-track imaging mode continuously adjusts the elevation and azimuth beam pointing of the satellite antenna to generate a mapping strip that matches the target terrain orientation. This makes it effective for certain terrain scenes that are not along the satellite track. This imaging mode achieves shorter observation time, better azimuth resolution, less invalid data, and higher data storage and processing efficiency than traditional modes, significantly improving the observation efficiency of spaceborne SAR for local terrain scenes that are not along the track orientation.

[0003] Differential InSAR (D-InSAR) is a technique that uses phase information from complex images of synthetic aperture radar to obtain information on surface changes, and it represents an extension of the applications of synthetic aperture radar satellites. Differential interferograms generated by satellite-generated synthetic aperture radar can be used to monitor Earth's surface deformation at the centimeter level or even smaller. Typically, topographic deformation information between two time-series images before and after deformation can be obtained through interferometric data generated from three flight passes (three-track method), or through differential interferometric processing of interferometric data generated from two flight passes and known topographic data of the observed area before deformation (two-track method).

[0004] Current spaceborne differential interferometric deformation measurements are mostly based on SAR images obtained from existing spaceborne frontal or small-slant-view strip observation modes. Traditional methods for designing multiple-track differential interferometric observations do not consider beam pointing maneuvers and lack the capability to map deformation in terrain scenes outside the satellite track. However, in non-track imaging modes, the advantages of beam spatial variation and high observation efficiency can be utilized to design multiple-track differential interferometric deformation measurement configurations. Currently, there is no research on differential interferometric measurements of local terrain deformation before and after non-track localization using multi-track SAR images acquired through non-track imaging observation modes, nor is there prior non-track regional terrain DEMs as auxiliary data. Therefore, it is necessary to study and expand the differential interferometric deformation measurement capabilities in non-track modes and further investigate the challenges of InSAR processing in non-track imaging observation modes. Summary of the Invention

[0005] This invention addresses the issue that in non-track imaging modes, the resulting interferometric baselines may suffer from inconsistencies in beam angles leading to interferometric decoherence, due to factors such as oblique-view geometry, double-track baseline parameters, spatial variations in the observation scene, and spatial variations in satellite antenna beam pointing. This makes it impossible to directly apply the differential interferometric processing methods of traditional spaceborne SAR front-side-view imaging modes for terrain deformation measurement. The invention provides a differential interferometric deformation measurement method for spaceborne SAR non-track modes, which can achieve the task of double-track interferometric processing of curved imaging zones in spaceborne non-track modes.

[0006] The spaceborne SAR non-track mode differential interferometric deformation measurement method of the present invention includes:

[0007] Step 1: Based on the non-track terrain area to be measured, determine the maximum and minimum values ​​of the downward viewing angle, the scene oblique angle, and the non-track observation angle during the observation process;

[0008] Step 2: Plan the three-track revisit baseline and design the beam pointing under the three-track revisit configuration; wherein, the first track is the main track; the second and third tracks are the interferometric revisit tracks of the first track. The second track constitutes the long interferometric baseline for measuring elevation, and the third track constitutes the short interferometric baseline for measuring deformation information after deformation; the design of the observation beam pointing variation of the second and third tracks is consistent with the design of the observation beam pointing variation of the first track to ensure that the difference in Doppler center frequency between interferometric image pairs is minimized;

[0009] Step 3: Based on the time sequence before and after deformation, select the non-track region observation data before deformation of the first and second tracks, and the non-track region observation data after deformation of the third track, respectively, to perform non-track time domain imaging, and obtain the main image, slave image I and slave image II.

[0010] Step 4: Combine the main image and secondary image I into a topographic image pair, and the main image and secondary image II into a deformation image pair. Perform pre-filtering and geometric registration processing on the two image pairs respectively.

[0011] Step 5: After performing interferometric phase generation and flatland phase processing on the terrain image pair and the deformation image pair respectively, phase filtering is then performed on each pair. Finally, phase unwrapping is performed to recover the wrapped phase, resulting in two pairs of interferometric image differential generation deformation phase maps.

[0012] Step 6: Based on system parameters, perform phase-to-deformation information and geocoding processing on the differential deformation phase maps generated from the two pairs of interferometric images to obtain surface deformation information of the non-track area.

[0013] Preferably, in step 2, a ground reference point is first selected in the direction of extension of the non-track topographic region to be measured, based on the wave foot direction, to generate the main orbit baseline; then, the consistency of the beam pointing variation range of the two double-track observations and the Doppler center frequency of the main orbit topographic observation range is controlled to ensure that the overlapping area covered by the wave positions of the second and third track non-track observations meets the requirements of interferometric measurement capability and coherence constraints, with the second track selecting a long interferometric baseline and the third track selecting a short interferometric baseline.

[0014] Ideally, the system bandwidth B, target slant range R, and beam angle should be considered first. Calculation of double orbit interference using beam-down angle θ, terrain slope η, wavelength λ, and speed of light c

[0015] Vertical critical baseline B ⊥crit as follows:

[0016]

[0017] The spatial baseline decoherence coefficient is then:

[0018]

[0019]

[0020] Among them, B ⊥ B is the vertical baseline. ⊥crit For the vertical critical baseline, ρ r For range resolution;

[0021] The azimuth Doppler center decoherence coefficient is:

[0022]

[0023] Where, ρ a dφ is the azimuth resolution, dφ is the difference in rotational viewing angle of the target scene between the two passing azimuth beams, and Δf is the azimuth resolution. DC To interfere with the azimuth-directed Doppler frequency shift of the data, W Dop For azimuth Doppler spatial spectral width;

[0024] The system signal-to-noise ratio (SNR) for decoherence is:

[0025]

[0026] Wherein, SNR represents the system signal-to-noise ratio:

[0027] The total coherence coefficient is then...

[0028] γ total =γ baseline ·γ D ·γ SNR

[0029] The interference baselines of the second and third tracks are selected based on the total coherence coefficient; the second track is selected with a long interference baseline, and the third track is selected with a short interference baseline.

[0030] Ideally, the effective baseline of the heavy rail should be controlled within 10-20% of the critical baseline constraint range.

[0031] Preferably, in step 3, non-track imaging is performed using a time-domain phase-preserving backward projection algorithm.

[0032] Preferably, in step 4, pre-filtering is performed on the two sets of image pairs respectively, and the pre-filtering is as follows:

[0033] First, calculate the spectral offsets in the range and azimuth directions under double orbit observations:

[0034]

[0035]

[0036] Where, Δf r , Δf a These are the range spectral offset and the azimuth spectral shift, respectively; η r f represents the slope of the terrain in the direction of distance; d1 and f d2 These are the azimuth center frequencies of the two-track imaging images, respectively; v g Let η be the wave foot velocity. a Orientation to terrain;

[0037] Then, by spectral truncation, non-common spectra are filtered out, and the common intervals of the spectrum are retained, thus completing the pre-filtering.

[0038] Preferably, in step 4, geometric registration processing is performed on the two sets of image pairs respectively. The geometric registration processing includes coarse geometric registration, pixel-level registration, and sub-pixel-level registration. Specifically, coarse registration involves: firstly, interpolating the orbital trajectory of the main satellite using ephemeris data, selecting control points in the main image, and obtaining their geographic coordinates through forward geocoding; then, obtaining the orbital trajectory of the satellite in the secondary image using the same fitting method, and obtaining its image coordinates in the secondary image through backward geocoding of the control points; then, subtracting the image coordinates in the two images yields the offsets in the azimuth and range directions, and compensating for the offsets between points in the main and secondary images based on these offsets.

[0039] Pixel-level registration is specifically as follows: a target region is selected in the main image, the coherence coefficient is calculated from the image to determine corresponding points, and a polynomial fitting is performed based on the determined corresponding points to obtain the offset distribution of the entire image, thereby achieving pixel-level registration.

[0040] Subpixel-level registration is specifically performed as follows: after completing pixel-level registration, non-baseband bilinear interpolation is first performed on the master and slave images. Then, the coherence coefficient is calculated using a sliding window, and the offset is fitted to achieve subpixel-level registration.

[0041] Preferably, step 5 specifically includes:

[0042] First, the interferometric phase maps of the two image pairs are generated separately; the expression for generating the interferometric phase by conjugate multiplication is as follows:

[0043]

[0044] Im1 and Im2 represent the master image and slave image in the image pair, respectively. int For the generated interferometric phase map, R is the slant range of the principal image, and ΔR is the difference between the slant ranges of the principal and slave images;

[0045] Then, flat-ground phase processing is performed: the slant range difference ΔR between the heavy rail antenna position and the scene target point is calculated, and the flat-ground phase of the reference plane is generated.

[0046]

[0047] The influence of the flat-ground phase can be removed by multiplying the flat-ground phase of the entire image with the original interferogram.

[0048] Next, phase filtering is performed on the image after removing the phase effects of flat terrain;

[0049] Finally, the minimum cost flow method is used for phase unwinding, and the absolute phase is determined based on the selected ground control points in the non-track region to obtain the unwinding phase result.

[0050] A better approach is to use a spatial median filtering algorithm to perform phase filtering on the image after removing the phase effects of flat terrain.

[0051] Preferably, in step 6, the phase-to-terrain calculation formula is as follows:

[0052]

[0053] μ def =B 2⊥ / B 1⊥

[0054] Where, Δd r The topographic deformation occurring along the line of sight. For deformation phase, The topographic phase before deformation, For the deformed terrain phase, μ def The effective baseline ratio between the deformation interferometer pair and the topographic interferometer pair;

[0055] The terrain deformation information obtained through phase-to-terrain calculation inversion is located in the radar image coordinate system (a, r, h). Then, the terrain deformation information in the image coordinate system is transformed to the geographic coordinate system through geocoding. The deformation coordinates of the terrain are used to obtain deformation measurement information.

[0056] Beneficial effects:

[0057] This invention establishes a double-track interferometric imaging processing model for non-track scenarios, enabling interferometric phase processing and terrain measurement inversion of local terrain. It primarily addresses the problem of terrain deformation measurement using the double-track interferometric capability based on the three-track measurement method in spaceborne non-track SAR configurations, proposing a design scheme to fill the gap in local terrain deformation measurement capabilities in non-track scenarios. Attached Figure Description

[0058] Figure 1 This is a flowchart of the measurement method of the present invention;

[0059] Figure 2 This is a schematic diagram illustrating the measurement principle of the present invention.

[0060] Figure 3 This is a schematic diagram of the measurement space configuration of the present invention. Detailed Implementation

[0061] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0062] This invention provides a method for measuring spaceborne SAR non-track mode differential interferometric deformation, such as... Figure 1 As shown, the specific steps include:

[0063] Step 1: Plan the configuration of the non-track imaging deformation measurement region;

[0064] Non-track imaging mode is used for observation of terrain scenes with curved extension characteristics and oblique tilt angles. In non-track imaging mode, the beam pointing in the elevation and azimuth dimensions is continuously adjusted to make the observation imaging zone scan along the target terrain direction.

[0065] Space configuration for spaceborne SAR non-track differential interferometry deformation measurement based on non-track imaging mode, such as... Figure 3 As shown, a non-track-based terrain region is selected based on the differential interferometric deformation measurement task. Where β f β n These represent the maximum and minimum downward angles during the observation process, respectively; H is the orbital altitude; Ω is the angle between the satellite track and the scene, called the scene oblique angle; W a W r These are the length and width of the diagonal scene, R. cα 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.

[0066] Step 2: Based on the satellite orbital parameters and the range of interferometry capability, plan the three-track revisit baseline. The first two track baselines are used for interferometric topographic surveying, and the third track baseline is used to obtain the topographic phase after deformation. The wavefront pointing under the three-track revisit configuration is also designed.

[0067] When designing the double orbit configuration and wave pointing, the first step is to calculate the movement trajectory and velocity parameters of the surface wave center based on the imaging capability of the first main orbit. Then, based on the satellite orbit parameters and the double orbit capability range, the two sets of double orbit baselines before and after deformation are designed to form the observation configuration of the terrain pair and deformation pair in the non-track mode.

[0068] The principle of differential interferometric deformation measurement using the three-track method in non-track mode is as follows: Figure 2 As shown, after determining the non-track topography to be observed, two sets of double orbit spatial baselines before and after deformation are planned based on satellite orbit parameters. Ground reference points are selected by determining the extension direction of the observation zone in the curved scene based on the wave foot direction. Constraints are applied to ensure the consistency of the Doppler center frequency between the beam pointing variation range of the two double orbit observations and the range of the first topographic observation, guaranteeing that the overlapping area covered by the wave positions of the two non-track double orbit observations meets the requirements of interferometric measurement capability and coherence constraints.

[0069] In the non-track imaging mode with slant view configuration, based on the system bandwidth B, target slant range R, and beam slant angle... Given the beam angle θ and terrain slope η, the critical baseline for heavy-track interferometry can be calculated as follows:

[0070]

[0071] Based on the satellite's non-track mode reorbit interferometry capability, the focus is mainly on three decoherence sources: spatial baseline decoherence, azimuth Doppler center decoherence, and system signal-to-noise ratio decoherence. The corresponding spatial baseline decoherence can be expressed as follows:

[0072]

[0073]

[0074] Among them B ⊥ B is the vertical baseline. ⊥crit For the vertical critical baseline, ρ r This represents the range resolution.

[0075] The decoherence calculation of the azimuth Doppler center is expressed as follows:

[0076]

[0077] Where θ is the downward angle, ρ a dφ is the azimuth resolution, dφ is the difference in rotational viewing angle of the target scene between the two passing azimuth beams, and Δf is the azimuth resolution. DC To interfere with the azimuth-directed Doppler frequency shift of the data, W Dop This represents the azimuth Doppler spatial spectral width.

[0078] The decoherent representation of the system signal-to-noise ratio is as follows, where SNR represents the system signal-to-noise ratio:

[0079]

[0080] The total coherence coefficient is expressed according to the above heavy rail correlation coefficient criterion as follows:

[0081] γ total =γ baseline ·γ D ·γ SNR

[0082] Where γ baseline and γ D Both are related to the change in beam angle of the observation area due to the master and slave flight paths, γ SNR The system signal-to-noise ratio depends on the non-track imaging platform. According to the coherence criterion shown in the above formula, the double-track interferometric baseline of the secondary overpass is preferred. To ensure overall coherence quality and terrain inversion accuracy, the effective double-track baseline is kept within 10-20% of the critical baseline constraint range. Figure 3 As shown, after terrain deformation occurs, the preferred third orbit track forms a shorter interference baseline with the original main track to create a deformation interference pair. The smaller effective deformation baseline B is then selected. 2⊥ Deformation interference pairs and larger effective baseline B 1⊥ Differential interferometry is performed on topographic interferometric pairs (controlled within the critical baseline constraint range). The effective baseline ratio of the deformation interferometric pair to the topographic interferometric pair is expressed as: Effective Baseline Ratio of Deformation Interferometric Pair to Topographic Interferometric Pair

[0083] μ def =B 2⊥ / B 1⊥

[0084] Topographic interferometry uses a larger baseline to ensure topographic measurement accuracy, while deformation interferometry uses a smaller baseline to ensure that the quality of topographic deformation information along the line of sight is less affected by terrain. The optimal baseline ratio for the interferometry is selected by considering the satellite's reorbiting capability. To ensure that no deformation occurs between the two time phases acquired by the topographic interferometry, it is necessary to optimize the temporal baseline of the two-track non-track topographic interferometry images acquired before deformation occurs, keeping it as short as possible.

[0085] After completing the baseline configuration selection, the first track with completed track planning is used as the master track. Based on the double track interferometry correlation coefficient criterion, the observation sampling positions of the two slave tracks are traversed, and the position with the smallest difference in beam pointing angle parameters after combining with the master track is selected to maintain the consistency of the observation beam pointing design.

[0086] The specific steps for wave position design selection are summarized as follows:

[0087] 1. Select a reference orbit. Using the non-track imaging orbit interval of a certain day as the reference orbit, record the aperture center time position, synthetic aperture time, Doppler bandwidth frequency, and beam center position parameters corresponding to the observed target;

[0088] 2. Select the heavy rail track baseline based on the reference track. Treat the reference track as the main track and the heavy rail track as the secondary track. Estimate the heavy rail baseline range based on the track interferometry correlation coefficient criterion and select a suitable secondary track baseline range.

[0089] 3. Select the aperture center position of the slave track based on the observation beam center position of the main track. After knowing the target-aperture center corresponding position and range of the main track, traverse all observation positions of the slave tracks, and select the matching aperture center time position in the slave track according to the criterion of minimizing the difference in beam pointing angle parameters between the double track position and the target point.

[0090] Based on the above, the differential interferometric baseline configuration and wave pointing design are completed.

[0091] Step 3: Based on the time sequence before and after deformation, select the topographic pairs of the two tracks before deformation and the observation data of the non-track area of ​​the one track after deformation, and apply the non-track time domain imaging algorithm for phase-preserving imaging processing.

[0092] Due to the spatial variation and two-dimensional coupling problem in spaceborne SAR non-track imaging mode observation, a non-track imaging time-domain imaging algorithm can be adopted, which has strong adaptability to spatial and time-varying parameters.

[0093] Non-track imaging temporal domain imaging algorithms are based on the phase-preserving back-projection (BP) algorithm. The basic idea of ​​this method is to back-project radar echo data onto each pixel of the imaging area, and then coherently superimpose the echoes at each pixel. The BP algorithm can complete imaging under any geometric configuration, without special requirements on the radar platform's orbital track or observation geometry. It is used as a reference algorithm in non-track configuration imaging methods.

[0094] The imaging process first establishes an oblique grid covering the entire scene along the target scene's extension direction. 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 temporal imaging efficiency of non-track imaging.

[0095] During the imaging process, a refined beamforming method is employed to address the issues of drastic Doppler center changes and spectral aliasing caused by two-dimensional beam scanning in non-track imaging. The imaging grid coordinates are transformed from the scene coordinate system XYZ to the satellite antenna coordinate system X'Y'Z'. The location of the beam-illuminated target is used to determine whether the imaging grid is illuminated. Based on the beam illumination determination result, the echo is then back-projected onto the 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.

[0096]

[0097] In back projection imaging, the Doppler phase of the target is removed before coherent superposition. After image reconstruction, the Doppler phase needs to be recovered based on the slant distance from the center of the scene observation zone. This involves phase preservation processing after coherent superposition in the azimuth direction of BP imaging. The slant distance R from the aperture center to the target reference point is selected. ref Phase compensation is performed on each image pixel value signal, and the phase-preserving processing is as follows:

[0098]

[0099] Where λ is the carrier frequency wavelength, which is the scene image obtained after phase preservation in non-track time-domain imaging.

[0100] Step 4: Select one of the two images before and after deformation as the main image, and the other two as secondary images. Perform pre-filtering and geometric registration processing on the secondary images and the main image respectively.

[0101] The principle of differential interferometry based on non-track images obtained from three-track measurement method is as follows: Figure 2 First, based on the squint geometry and baseline parameters of the non-track imaging mode, pre-filtering and geometric registration are performed on the two pairs of double-track images:

[0102] (1) Pre-filtering of heavy rail images

[0103] When observing the same target via multiple orbits in non-track imaging mode, inconsistencies in beam angle, baseline, and Doppler parameters can lead to spectral shifts in the multiple orbit images. Therefore, a pre-filtering step is required before generating the interferometric phase of the multiple orbit images for terrain and deformation pairs. The range-direction spectral shift Δf under multiple orbit observations... rand azimuth spectral offset Δf a They are represented as follows:

[0104]

[0105]

[0106] Where 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 This represents the azimuth slope of the terrain. The spectral offset can be directly calculated from the system parameters. Through spectral truncation, non-common spectrum is filtered out, leaving only the common intervals of the retained spectrum.

[0107] (2) Geometric registration of the two sets of images

[0108] Two sets of non-track imaging master and slave images of the double orbit constitute the terrain pair and deformation pair, respectively. Geometric coarse registration is performed on the two pairs of double orbit images, and three-level spatial registration at the pixel and sub-pixel level is performed using the correlation function method.

[0109] coarse registration

[0110] First, coarse geometric registration is performed between the master and slave images based on their imaging geometry. The master satellite's trajectory is obtained by interpolation using its ephemeris data, and control points in the master image are selected and geocoded forward to obtain their geographic coordinates. Then, the slave image's satellite trajectory is obtained using the same fitting method, and the control points are geocoded backward to obtain their image coordinates in the slave image. Finally, subtracting the image coordinates from the two images yields the azimuth and range offsets. These offsets are then corrected to complete the coarse registration.

[0111] Pixel-level registration

[0112] Starting from the internal information of the master and slave images, several control points and corresponding matching windows are selected in the master image; corresponding search windows are selected in the slave image. By sliding the matching window in the search window, the corresponding offset is selected according to the real correlation function (quasi-measure function), and two-dimensional fitting is performed to fit the two-dimensional surfaces of the new coordinates in the range and azimuth directions respectively. The slave image is then resampled to complete pixel-level registration.

[0113] Subpixel registration

[0114] Subpixel-level registration is then performed. In non-track imaging, when performing azimuth interpolation resampling from the image, the Doppler center frequency f at each point in the azimuth direction must be considered. dc Unlike other methods, the Doppler center frequency f is considered for the pixels to be interpolated in the scene. dc The baseband interpolation is converted to non-baseband interpolation, and bilinear interpolation is used for interpolation resampling. After performing non-baseband bilinear interpolation on the master and slave images, a sliding window is used to calculate the coherence coefficients, and offset fitting is performed to complete sub-pixel level registration.

[0115] Step 5: Perform interference phase generation and flat-ground phase processing on the two sets of images before and after deformation, then perform phase filtering processing on each, and finally perform phase unwrapping processing to restore the wrapped phase.

[0116] After initial registration, the master and slave images of the two groups are processed to generate interferometric phase maps, followed by flat-ground effect removal. The interferometric phase expression generated by conjugate multiplication is as follows:

[0117]

[0118] Two of the images are Im1 and Im2, 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.

[0119] The acquired orbital track data and heavy orbit baseline parameters are used for de-flattening. Given precise orbital ephemeris data, the slant range difference ΔR between the heavy orbit antenna position and the target point in the scene is calculated, and the flat-ground phase of the reference plane is generated. The reference flatland phase term is represented as follows:

[0120]

[0121] The influence of the flat-ground phase can be removed by multiplying the flat-ground phase with the original interferogram after generating the flat-ground phase of the whole image.

[0122] After removing the flat-ground phase processing, the spatial median filtering algorithm is used to filter out noise in the interferogram after removing the flat-ground effect phase filtering.

[0123] After phase filtering, the minimum cost flow (MCF) method is used to unwrap the denoised interference phase, and the absolute phase is determined by the selected ground control points in the non-track region to obtain the unwrapped phase result.

[0124] Step 6: Generate a deformation phase map by differential analysis of the two pairs of interferometric images after the above interferometric processing. Based on the system parameters, perform phase-to-deformation information and geocoding processing to obtain the surface deformation information of the non-track area.

[0125] Deformation phases are generated by differentiating two pairs of interferometric topographic images after interferometry processing. These deformation phases are then subjected to topographic inversion and geocoding. The phase-to-topographic calculation formula is as follows:

[0126]

[0127] μ def =B 2⊥ / B 1⊥

[0128] Where, Δd r The topographic deformation occurring along the line of sight. For deformation phase, The topographic phase before deformation, For the deformed terrain phase, μ def This represents the effective baseline ratio between the deformation interferometric pair and the terrain interferometric pair. The terrain deformation information obtained through phase-to-terrain calculation inversion is located in the radar image coordinate system (a, r, h). Then, geocoding is used to transform the terrain deformation information in the image coordinate system to the geographic coordinate system. The terrain deformation coordinates below.

[0129] Ultimately, this yielded topographic deformation measurement information for the two non-track observation areas before and after deformation.

[0130] Therefore, this invention provides a spaceborne SAR non-track mode differential interferometric deformation measurement method, which makes up for the existing non-track terrain deformation measurement information acquisition capabilities.

[0131] 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 method for measuring differential interferometric deformation in a non-track mode of spaceborne SAR, characterized in that, include: Step 1: Based on the non-track terrain area to be measured, determine the maximum and minimum values ​​of the downward viewing angle, the scene oblique angle, and the non-track observation angle during the observation process; Step 2: Plan the three-track revisit baseline and design the beam pointing under the three-track revisit configuration; wherein, the first track is the main track; the second and third tracks are the interferometric revisit tracks of the first track. The second track constitutes the long interferometric baseline for measuring elevation, and the third track constitutes the short interferometric baseline for measuring deformation information after deformation; the design of the observation beam pointing variation of the second and third tracks is consistent with the design of the observation beam pointing variation of the first track to ensure that the difference in Doppler center frequency between interferometric image pairs is minimized; In step 2, firstly, based on the wave foot direction, a ground reference point is selected in the extension direction of the non-track topographic region to be measured, generating the main orbit baseline; then, the consistency of the beam pointing variation range of the two repeated orbit observations and the Doppler center frequency of the main orbit topographic observation range is controlled to ensure that the overlapping area covered by the wave positions of the second and third track non-track observations meets the requirements of interferometric measurement capability and coherence constraints. A long interferometric baseline is selected for the second track, and a short interferometric baseline is selected for the third track; specifically... First, based on the system bandwidth B, target slant range R, and beam angle... Calculate the vertical critical baseline B for heavy-orbit interferometry using the following parameters: beam angle θ, terrain slope η, wavelength λ, and speed of light c. ⊥crit as follows: The spatial baseline decoherence coefficient is then: Among them, B ⊥ B is the vertical baseline. ⊥crit For the vertical critical baseline, ρ r For range resolution; The azimuth Doppler center decoherence coefficient is: Where, ρ a dφ is the azimuth resolution, dφ is the difference in rotational viewing angle of the target scene between the two passing azimuth beams, and Δf is the azimuth resolution. DC To interfere with the azimuth-directed Doppler frequency shift of the data, W Dop For azimuth Doppler spatial spectral width; The system signal-to-noise ratio (SNR) for decoherence is: Wherein, SNR represents the system signal-to-noise ratio: The total coherence coefficient is then... c total =c baseline ·c D ·c SNR The double-track interference baselines for the second and third tracks are selected based on the total coherence coefficient; among them, the second track is selected with a long interference baseline, and the third track is selected with a short interference baseline. Step 3: Based on the time sequence before and after deformation, select the non-track region observation data before deformation of the first and second tracks, and the non-track region observation data after deformation of the third track, respectively, to perform non-track time domain imaging, and obtain the main image, slave image I and slave image II. Step 4: Combine the main image and secondary image I into a topographic image pair, and the main image and secondary image II into a deformation image pair. Perform pre-filtering and geometric registration processing on the two image pairs respectively. Step 5: After performing interferometric phase generation and flatland phase processing on the terrain image pair and the deformation image pair respectively, phase filtering is then performed on each pair. Finally, phase unwrapping is performed to recover the wrapped phase, resulting in two pairs of interferometric image differential generation deformation phase maps. Step 6: Based on system parameters, perform phase-to-deformation information and geocoding processing on the differential deformation phase maps generated from the two pairs of interferometric images to obtain surface deformation information of the non-track area.

2. The method as described in claim 1, characterized in that, The effective baseline of the heavy rail should be controlled within 10-20% of the critical baseline constraint range.

3. The method as described in claim 1, characterized in that, In step 3, non-track imaging is performed using a time-domain phase-preserving backward projection algorithm.

4. The method as described in claim 1, characterized in that, In step 4, pre-filtering is performed on the two sets of image pairs respectively. The pre-filtering is as follows: First, calculate the spectral offsets in the range and azimuth directions under double orbit observations: Where, Δf r , Δf a These are the range spectral offset and the azimuth spectral shift, respectively; η r f represents the slope of the terrain in the direction of distance; d1 and f d2 These are the azimuth center frequencies of the two-track imaging images, respectively; v g Let η be the wave foot velocity. a Orientation to terrain; Then, by spectral truncation, non-common spectra are filtered out, and the common intervals of the spectrum are retained, thus completing the pre-filtering.

5. The method as described in claim 1, characterized in that, In step 4, geometric registration is performed on the two sets of image pairs. The geometric registration process includes coarse geometric registration, pixel-level registration, and sub-pixel-level registration. Specifically, coarse registration involves: firstly, interpolating the orbital trajectory of the main satellite using ephemeris data, selecting control points in the main image, and obtaining their geographic coordinates through forward geocoding; then, obtaining the orbital trajectory of the satellite in the secondary image using the same fitting method, and obtaining its image coordinates in the secondary image through backward geocoding of the control points; finally, subtracting the image coordinates in the two images yields the offsets in the azimuth and range directions, and the offsets between points in the main and secondary images are compensated based on these offsets. Pixel-level registration is specifically as follows: a target region is selected in the main image, the coherence coefficient is calculated from the image to determine corresponding points, and a polynomial fitting is performed based on the determined corresponding points to obtain the offset distribution of the entire image, thereby achieving pixel-level registration. Subpixel-level registration is specifically performed as follows: after completing pixel-level registration, non-baseband bilinear interpolation is first performed on the master and slave images. Then, the coherence coefficient is calculated using a sliding window, and the offset is fitted to achieve subpixel-level registration.

6. The method as described in claim 1, characterized in that, Step 5 specifically involves: First, the interferometric phase maps of the two image pairs are generated separately; the expression for generating the interferometric phase by conjugate multiplication is as follows: Im1 and Im2 represent the master image and slave image in the image pair, respectively. int For the generated interferometric phase map, R is the slant range of the principal image, and ΔR is the difference between the slant ranges of the principal and slave images; Then, flat-ground phase processing is performed: the slant range difference ΔR between the heavy rail antenna position and the scene target point is calculated, and the flat-ground phase of the reference plane is generated. The influence of the flat-ground phase can be removed by multiplying the flat-ground phase of the entire image with the original interferogram. Next, phase filtering is performed on the image after removing the phase effects of flat terrain; Finally, the minimum cost flow method is used for phase unwinding, and the absolute phase is determined based on the selected ground control points in the non-track region to obtain the unwinding phase result.

7. The method as described in claim 6, characterized in that, A spatial median filtering algorithm is used to perform phase filtering on the image after removing the phase effects of flat terrain.

8. The method as described in claim 1, characterized in that, In step 6, the phase-to-terrain calculation formula is as follows: Where, Δd r The topographic deformation occurring along the line of sight. For deformation phase, The topographic phase before deformation, For the deformed terrain phase, μ def The effective baseline ratio between the deformation interferometer pair and the topographic interferometer pair; The terrain deformation information obtained through phase-to-terrain calculation inversion is located in the radar image coordinate system (a,r,h). Then, geocoding is used to transform the terrain deformation information in the image coordinate system to the geographic coordinate system. The deformation coordinates of the terrain are used to obtain deformation measurement information.

Images