Three-dimensional flow velocity field inversion method and system based on coherence-assisted SAR adaptive offset tracking, and storage medium

Abstract

The invention provides a three-dimensional flow velocity field inversion method and system based on coherence-assisted SAR adaptive offset tracking, and a storage medium, so as to improve the problem of decoherence caused by time heterogeneity in a synthetic aperture radar interferometry (InSAR) technology, and relates to the technical field of electric power engineering geological disaster monitoring. The method comprises the following steps: establishing a space-time baseline network between SAR images; obtaining coherence data after geocoding; selecting a pair of SAR images from the space-time baseline network as a reference unit, calculating an offset result through a pixel offset tracking technology for the reference unit, and determining an optimal matching window parameter of the reference unit; establishing a mapping function between coherence and optimal matching window parameters; according to the established mapping function, calculating an optimal matching window parameter of each remaining SAR image pair; and in combination with the line-of-sight average deformation rate fields of the ascending and descending rails, inverting a surface three-dimensional flow velocity deformation field of the target area through a geometric vector decomposition model.

Description

Technical Field

[0001] This invention relates to the field of power engineering geological disaster monitoring technology, specifically a three-dimensional velocity field inversion method, system, and storage medium based on coherence-assisted SAR adaptive offset tracking. Background Technology

[0002] In the operation and maintenance of power engineering projects, geological disasters such as landslides and glacial disturbances pose a serious threat to critical facilities such as transmission lines and substations. Transmission lines often traverse complex terrain spanning hundreds of kilometers, with varying geological conditions along the route, making it difficult for traditional manual inspection methods to achieve comprehensive coverage and timely early warning. Especially in seismically active areas, mining subsidence areas, and landslide-prone areas, surface deformation can lead to serious consequences such as tower tilting and foundation displacement, directly affecting the safe and stable operation of the power grid.

[0003] Synthetic Aperture Radar (InSAR) technology, with its wide-area and high-precision monitoring capabilities, offers a new solution for monitoring deformation in power facilities. Compared to traditional measurement methods, InSAR technology can achieve: (1) Regularly scan wide-area power facilities, with a monitoring range of tens of thousands of square kilometers; (2) Penetrate through cloud and rain weather to ensure the continuity of monitoring data; (3) Identify surface deformation at the millimeter to centimeter level to meet the safety early warning requirements of power facilities.

[0004] The technical approaches for measuring surface deformation using SAR data can be broadly categorized into two types: phase-based interferometry and intensity-based offset tracking. While InSAR technology boasts superior accuracy, its application is strictly limited by interferometric coherence. In practical applications, the diverse surface cover and complex surface features in areas with power facilities, especially dense vegetation, easily lead to decoherence in traditional InSAR technology. When InSAR fails due to coherence loss, pixel offset tracking (POT) becomes a mature and necessary supplementary technique. Offset tracking does not rely on phase information but directly utilizes the amplitude of SAR images. The main advantage of offset tracking lies in its insensitivity to phase decoherence and its ability to measure large gradient deformations on the order of meters or even tens of meters.

[0005] In offset tracking algorithms, the matching window size is a core parameter determining the accuracy and reliability of displacement field inversion, and its setting faces an inherent trade-off between signal-to-noise ratio (SNR) and local deformation fidelity. On the one hand, a larger matching window contains rich surface texture features, which can effectively suppress speckle noise by increasing the statistical sample size, thereby significantly improving the SNR and matching robustness of the cross-correlation peak. On the other hand, an excessively large window can introduce a severe spatial averaging effect, causing local non-uniform deformation information to be masked. Conversely, a smaller window, while reducing smoothing errors, is highly susceptible to noise interference, leading to matching failure. Therefore, the standard practice of using a single fixed-size window throughout the study area is essentially a compromise. Although this strategy performs reasonably well in scenarios with uniform surface features and deformation patterns, its limitations are particularly prominent in strongly heterogeneous environments: spatially, for long-distance transmission corridors spanning multiple regions, a fixed window cannot simultaneously account for the feature differences between undulating and flat areas; temporally, facing the seasonal changes in vegetation growth, a window based on static spatial features cannot adapt to the dynamic evolution of signal quality. Therefore, the traditional fixed window method can no longer meet the needs of power facility monitoring for refined adaptation to complex environments, and the development of adaptive window strategies is imperative.

[0006] While the necessity of adaptive windowing strategies in offset tracking is widely recognized, previous research has primarily focused on addressing spatial heterogeneity issues (such as landslide boundary preservation), and remains limited in handling temporal complexity. Furthermore, existing methods largely rely on indirect features such as image texture and predicted gradients, or are based on empirical rules, resulting in high computational costs and a lack of rigorous physical foundations. Essentially, these indicators lack a direct and quantitative correlation with the core physical process of SAR signal degradation over time—the temporal evolution of scattering characteristics. Therefore, given the significant seasonal variations and multi-temporal observation requirements in power facility monitoring, there is an urgent need to establish a physically driven and computationally efficient adaptive windowing method. This method should be able to start from first principles, intrinsically and quantitatively guiding the selection of window parameters through the temporal decay law of signal quality, thereby overcoming the monitoring challenges in complex time-varying environments. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide a three-dimensional velocity field inversion method, system and storage medium based on coherence-assisted SAR adaptive offset tracking, so as to improve the decoherence problem caused by temporal heterogeneity in synthetic aperture radar interferometry (InSAR) technology.

[0008] The technical solution adopted by this invention to solve its technical problem is: a three-dimensional velocity field inversion method based on coherence-assisted SAR adaptive offset tracking, comprising the following steps: S1. Collect SAR data and digital elevation model data covering the target area, preprocess them, and establish a spatiotemporal baseline network between SAR images. S2. Based on the spatiotemporal baseline network, perform differential interferometry on each pair of SAR images to obtain their geocoded coherent data. S3. Select a pair of SAR images from the spatiotemporal baseline network as a reference unit, and calculate the offset result for the reference unit using pixel offset tracking technology, and determine its optimal matching window parameters. S4. Based on the obtained geocoded coherence data of the reference unit and the determined best matching window parameters of the reference unit, establish a mapping function between coherence and best matching window parameters; S5. Based on the established mapping function, substitute the geocoded coherence data of the remaining unprocessed SAR image pairs in the spatiotemporal baseline network, calculate the optimal matching window parameters for each remaining SAR image pair, and thus complete the determination of the optimal matching window parameters for all SAR image pairs in the spatiotemporal baseline network. S6. Using the optimal matching window parameters of all determined SAR image pairs, perform pixel offset tracking processing on all SAR image pairs in the spatiotemporal baseline network to obtain the line-of-sight displacement field of each SAR image pair. S7. Based on the small baseline set technique, the line-of-sight displacement field of each obtained SAR image is calculated in time series to obtain the line-of-sight average deformation rate field of the ascending and descending orbits. S8. By combining the line-of-sight average deformation rate fields of the ascending and descending orbits, the three-dimensional surface velocity deformation field of the target area is inverted through a geometric vector decomposition model.

[0009] Furthermore, the preprocessing includes: Based on the projection equation, the digital elevation model data is cropped, resampled, and transformed in coordinate system using satellite orbit parameters to make its spatial range and reference system consistent with the SAR data. Select one SAR image as the super master image, and finely register the remaining SAR images to the radar coordinate system of the super master image; Based on preset temporal baseline thresholds and spatial vertical baseline thresholds, all registered SAR images are combined in pairs and filtered to construct the spatiotemporal baseline network.

[0010] Further, step S2 includes: S21. For each pair of SAR images in the spatiotemporal baseline network, determine the master image and slave image in chronological order, perform multi-view, registration, and interferometry on the two images to generate a differential interferogram, and use the digital elevation model data and orbital parameters to remove the terrain phase and flat land phase. S22. Filter the generated differential interferogram to obtain the filtered differential interferogram; and calculate the coherence coefficient of each pixel to obtain coherence data. S23. Based on satellite orbital parameters and digital elevation model data, the obtained filtered differential interferogram and coherence data are transformed from the radar coordinate system to the target geodetic coordinate system to complete geocoding and obtain geocoded coherence data.

[0011] Furthermore, the determination of the optimal matching window parameters using pixel offset tracking technology includes: On the image corresponding to the reference unit, a series of matching windows of different sizes are used to perform pixel offset tracking calculations. Compare the quality of displacement field results under different window sizes; Select the window size corresponding to the best quality assessment result as the optimal matching window parameter.

[0012] Furthermore, the mapping function is: ; In the formula, : Optimal matching window size; Coherence.

[0013] Furthermore, in step S5, the optimal matching window parameters for the SAR image pair to be processed are calculated using the following formula: ; In the formula, : The optimal offset tracking matching window for the reference cell; Reference unit coherence; : The best matching window for offset tracking of the SAR image pair to be processed; : Coherence of the SAR image pair to be processed.

[0014] Further, step S7 includes: S71. Construct a set of observation equations H·V=δ, with the line-of-sight displacement of each SAR image pair as the observed value and the average deformation rate of each adjacent time interval as the unknown quantity, where δ is the observation vector composed of the displacement values ​​of all SAR image pairs, V is the average deformation rate sequence vector to be determined, and H is the design matrix. S72. Solve the observation equations using the singular value decomposition method to obtain the line-of-sight average deformation rate fields for the ascending and descending orbits, respectively.

[0015] Furthermore, the geometric vector decomposition model is a constraint method based on the parallel flow hypothesis.

[0016] A three-dimensional velocity field inversion system based on coherence-assisted SAR adaptive migration tracking is used to execute the aforementioned three-dimensional velocity field inversion method based on coherence-assisted SAR adaptive migration tracking, including: The data preprocessing and network construction module is configured to perform step S1. The coherence calculation module is configured to execute step S2. An adaptive window decision module is configured to execute steps S3, S4, and S5. The displacement field extraction and time-series calculation module is configured to execute steps S6 and S7; and... The three-dimensional inversion module is configured to execute step S8.

[0017] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described three-dimensional velocity field inversion method based on coherence-assisted SAR adaptive offset tracking.

[0018] The beneficial effects of this invention are as follows: By establishing a physical model between coherence and offset tracking optimal window, this invention achieves adaptive and precise matching of processing parameters to the spatiotemporal changes of surface scattering characteristics. To a certain extent, it overcomes the limitations of fixed window methods in complex scenarios where "accuracy" and "robustness" are difficult to balance. It can simultaneously improve the spatial detail resolution of deformation monitoring results and the reliability in low coherence areas. By replacing "human experience trial and error" with "physical model-driven" approach, it achieves automation and standardization of the processing flow, effectively improving the accuracy and precision of surface deformation assessment and significantly increasing computational efficiency. This method effectively integrates the technical advantages of InSAR and POT, expanding the effective spatiotemporal monitoring range. It is particularly suitable for the safe operation and maintenance of major projects such as power facility corridors that span multiple landforms, require seasonal changes, and need long-term stable monitoring. It provides a better technical solution for geological disaster identification and infrastructure lifecycle safety management. Attached Figure Description

[0019] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0020] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0021] like Figure 1As shown, the three-dimensional velocity field inversion method based on coherence-assisted SAR adaptive offset tracking of the present invention includes the following steps: S1. Collect ascending-orbit synthetic aperture radar (SAR) data and digital elevation model data covering the target area, preprocess them, and establish a spatiotemporal baseline network between SAR images. S2. Based on the spatiotemporal baseline network, perform differential interferometry on each pair of SAR images to obtain their geocoded coherent data. S3. Select a pair of SAR images from the spatiotemporal baseline network as a reference unit, and determine the optimal matching window parameters for the reference unit using pixel offset tracking technology. S4. Based on the obtained geocoded coherence data of the reference unit and the determined best matching window parameters of the reference unit, establish a mapping function between coherence and best matching window parameters; S5. Based on the established mapping function, substitute the geocoded coherence data of the remaining unprocessed SAR image pairs in the spatiotemporal baseline network, calculate the optimal matching window parameters for each remaining SAR image pair, and thus complete the determination of the optimal matching window parameters for all SAR image pairs in the spatiotemporal baseline network. S6. Using the optimal matching window parameters of all determined SAR image pairs, perform pixel offset tracking processing on all SAR image pairs in the spatiotemporal baseline network to obtain the line-of-sight displacement field of each SAR image pair. S7. Based on the small baseline set technique, the line-of-sight displacement field of each obtained SAR image is calculated in time series to obtain the line-of-sight average deformation rate field of the ascending and descending orbits. S8. By combining the line-of-sight average deformation rate fields of the ascending and descending orbits, the three-dimensional surface velocity deformation field of the target area is inverted through a geometric vector decomposition model.

[0022] This invention establishes a physical model between coherence and offset tracking optimal windows, achieving adaptive and precise matching of processing parameters to spatiotemporal variations in surface scattering characteristics. This overcomes, to some extent, the core contradiction of the fixed-window method—the difficulty in balancing accuracy and robustness in complex scenarios—and simultaneously improves the spatial detail resolution and reliability of deformation monitoring results in low-coherence regions. By replacing "trial and error based on human experience" with a "physical model-driven" approach, the processing workflow is automated and standardized, effectively improving the accuracy and precision of surface deformation assessment and significantly increasing computational efficiency. This method effectively integrates the technological advantages of InSAR and POT, expanding the effective spatiotemporal monitoring range. It is particularly suitable for the safe operation and maintenance of major projects such as power facility corridors that span various terrains, require seasonal changes, and need long-term stable monitoring, providing a superior technical solution for geological disaster identification and infrastructure lifecycle safety management.

[0023] In this embodiment of the invention, the preprocessing includes: Based on the projection equation, the digital elevation model data is cropped, resampled, and transformed in coordinate system using satellite orbit parameters to ensure that its spatial range and reference system are consistent with the SAR data. Select one SAR image as the super master image, and finely register the remaining SAR images to the radar coordinate system of the super master image; Based on preset temporal baseline thresholds and spatial vertical baseline thresholds, all registered SAR images are combined in pairs and filtered to construct the spatiotemporal baseline network.

[0024] The above preprocessing method ensures that the comparison of each pixel in subsequent steps (such as differential interferometry and offset tracking) is performed at the same ground position in a strict sense. This is the most fundamental prerequisite for obtaining high-precision deformation information and avoids false signals or accuracy loss caused by data misalignment.

[0025] In this embodiment of the invention, registering multiple SAR images to the radar coordinate system of the super master image includes: Super master image selection: Select one image with good imaging quality (meeting preset requirements) and centered spatiotemporal baseline from all SAR images as the super master image. Registration process: Register each of the remaining SAR images (secondary images) with the super master image one by one. Specifically, this includes: initial geometric correction based on satellite orbital parameters; on this basis, a high-precision registration model is constructed by automatically extracting a large number of corresponding feature points on the two images and calculating their offsets; finally, the secondary images are resampled according to the model to generate a new image that is completely aligned with the super master image on the pixel grid.

[0026] In this embodiment of the invention: step S2 includes: S21. For each pair of SAR images in the spatiotemporal baseline network, determine the master image and slave image in chronological order, perform multi-view, registration, and interferometry on the two images to generate a differential interferogram, and use the digital elevation model data and orbital parameters to remove the terrain phase and flat phase; in this embodiment of the invention, a multi-view ratio of 10:2 is set to reduce the influence of speckle noise. S22. Filter the generated differential interferogram (e.g., Goldstein filtering) to suppress phase noise and optimize interference fringes to obtain the filtered differential interferogram; and calculate the coherence coefficient of each pixel to obtain coherence data. S23. Based on satellite orbital parameters and digital elevation model data, the filtered differential interferogram obtained in S23 and the coherence data obtained in S22 are converted from radar coordinate system one to geographic coordinate system to complete geocoding and obtain geocoded coherence data.

[0027] In this embodiment of the invention, a pair of SAR images is randomly selected from the spatiotemporal baseline network as a reference unit.

[0028] In some embodiments, the image pair with the shortest combined spatial and temporal baselines (or each below a threshold) is selected as the reference unit.

[0029] In this invention, the specific process of establishing the mapping function between coherence and the optimal matching window parameters based on the calculated offset and the determined optimal matching window parameters of the reference unit is as follows: S31. Define the total variance of the offset estimation as... ;

[0030] in, It is a decoherence / noise error; It is deformation gradient error; S32. Based on interferometry theory, Modeling as coherence γ Inversely proportional to the window area Inversely proportional functions: ; in, A It is a proportionality constant that includes factors such as sensor noise level and signal brightness; S33. Establish deformation gradient error The model; when a deformation gradient exists within the matching window (i.e., the displacement is not constant but varies spatially), the "average displacement" that the offset tracking algorithm attempts to find is itself an approximation; the larger the window, the more drastic the displacement changes it contains, and the greater the error introduced by this approximation; assuming a linear deformation gradient exists in one direction. k (Unit: pixel displacement / pixel distance); in a width of W Within the window, the maximum displacement difference between the two ends is k · W The variance of the error between the estimated average displacement and the actual displacement field can be reasonably assumed to be proportional to the square of the maximum displacement difference. ; in, B It is a proportionality constant.

[0031] S34. Establish a unified model for total variance.

[0032] Adding the two error sources together gives the total variance with respect to the window size. W coherence γ and deformation gradient k Functions: ; in, A It is a constant; Therefore, increase W This will decrease the first term (noise error), but will increase the second term (gradient error); decreasing W It will decrease the second term (gradient error), but will increase the first term (noise error).

[0033] S35. Solving for the optimal window size

[0034] 1) To W Differentiate: ; 2) Set the derivative to zero: ; 3) Solve W : ; ; ; To simplify, the constant term can be... A , B and local surface properties kThey are merged into a new constant because it is relatively fixed for a specific deformation region; Therefore, we get: .

[0035] In this invention: For any other SAR image pair to be processed in the spatiotemporal baseline network, the average coherence is calculated based on its corresponding coherence coefficient map, and the optimal matching window parameters are adaptively determined by combining the constructed mapping function, as follows: ; In the formula, : The optimal offset tracking matching window for the reference cell; Reference unit coherence; : The best matching window for offset tracking of the SAR image pair to be processed; : Coherence of the SAR image pair to be processed.

[0036] The specific process of step S7 is as follows: S71. Construct a set of observation equations H·V=δ, where the line-of-sight displacement of each SAR image is the observed value and the average deformation rate of each adjacent time interval is the unknown quantity. Here, δ is the observation vector composed of the displacement values ​​of all SAR image pairs, V is the sequence vector of the average deformation rate to be calculated, and H is the design matrix. Assume there are N SAR images at N time points, and the objective is to calculate the average deformation rate vector V1, V2, ..., V1 over N-1 time intervals. N-1 For M SAR image pairs, their observed displacement vector δ and deformation rate vector V satisfy a linear relationship H·V=δ. Each row of the design matrix H corresponds to an image pair, and the element values ​​are determined by the start and end times of that image pair.

[0037] S72. Solve the observation equations using singular value decomposition (SVD) to obtain the line-of-sight average deformation rate fields for the ascending and descending orbits, respectively. Since the spatiotemporal baseline network is typically composed of multiple discontinuous time periods (subsets), matrix H is rank-deficient and cannot be solved using conventional least squares. Singular value decomposition (SVD) is employed: H = U∑V T U: An M×M orthogonal matrix whose column vectors span the observation space (i.e., the space containing δ). Each column of U is called a left singular vector. ∑: An M×(N-1) matrix whose elements are all 0 except for those on the main diagonal (called singular values). V: An (N-1)×(N-1) orthogonal matrix whose column vectors span the parameter space (i.e., the space containing V). Each column of V is called a right singular vector. By discarding singular values ​​close to zero, the stable minimum norm solution v=V∑ is obtained by finding the generalized inverse. + UTδ; v: the average deformation rate vector obtained from the solution; ∑ +The pseudo-inverse (generalized inverse) matrix of ∑ is obtained. By performing the above solution independently for each pixel, the time-series line-of-sight deformation rate fields of the ascending and descending orbits can be obtained.

[0038] Inversion can be achieved through geometric vector decomposition models, including but not limited to: joint solutions based on multi-source data fusion, constraint solutions based on specific physical models (such as parallel flows), or filtering estimation algorithms based on spatiotemporal priors. This embodiment preferably adopts a constraint method based on the parallel flow hypothesis.

[0039] In this embodiment of the invention, the specific process of inverting the three-dimensional surface velocity deformation field of the target area using the constraint method based on the parallel flow hypothesis in the geometric vector decomposition model is as follows: S81. Constructing the observation equation set: Obtain the InSAR line-of-sight average deformation rate field of the ascending and descending satellites; for each pixel in the target area, using its ascending and descending line-of-sight deformation rate observations, combined with satellite geometric parameters, establish a system of linear equations between its three-dimensional deformation rate (eastward, northward, and vertical) and the two line-of-sight observations: ; in: , These are the observed average deformation rates along the line of sight for the ascending and descending orbits at the same ground point; Let be the three-dimensional deformation rate vector of this ground point, where , , These represent the deformation rate components in the east, north, and vertical directions, respectively. and These are the direction cosine components of the unit vectors in the line-of-sight direction of the ascending and descending satellites in the local northeast-sky coordinate system, calculated from the satellite orbit geometry (incident angle, azimuth angle); , These represent the noise errors included in the ascending and descending orbit observations, respectively. S82. Introducing the parallel flow hypothesis constraint: Based on the topographic features of the target area or known physical models, it is assumed that surface deformation mainly manifests as quasi-two-dimensional flow along a certain dominant horizontal direction, that is, there is a fixed proportional relationship between the northward velocity component and the eastward velocity component. , For flow direction constraint coefficient, θ is the angle (azimuth) between the pre-defined dominant flow direction at that pixel and the due east direction; substituting this geometric constraint into the equation set in step S81 reduces the unknowns from three velocity components per pixel to two. S83. Construct a regional overall inversion matrix model: The constrained equations of all pixels in the target area are integrated into a unified matrix equation, expressed in linear form: d=Gm+e, where d is the observation vector composed of all ascending and descending orbit observations, m is the parameter vector composed of the velocity component parameters to be determined for all pixels, G is the coefficient matrix (design matrix) composed of satellite geometry and flow constraints, and e is the observation noise vector. S84. Truncation Singular Value Decomposition (TSVD) Solution: Perform singular value decomposition on the coefficient matrix G; by setting a truncation threshold, discard singular values ​​with excessively small values ​​and their corresponding singular vectors to overcome matrix ill-conditioning caused by poor observation geometry or insufficient constraints; use the truncated singular value decomposition results to calculate stable parameter vector estimates m.

[0040] S85. Reconstruction and Output of Three-Dimensional Velocity Field: Based on the parameter vector m obtained from step S84, the eastward and vertical deformation rates of each pixel are recovered; combined with the parallel flow constraint relationship introduced in step S82, the northward deformation rate is calculated; finally, the three-dimensional surface velocity deformation field with three components (eastward, northward, and vertical) covering the entire target area is synthesized and output.

Claims

1. A three-dimensional flow velocity field inversion method based on coherence-assisted SAR adaptive offset tracking, characterized in that, The method comprises the following steps: S1, collecting and preprocessing ascending track SAR (Synthetic Aperture Radar) data and digital elevation model data covering a target area to establish a time-space baseline network among SAR images; S2, based on the time-space baseline network, performing differential interference processing on each pair of SAR image pairs to obtain their geocoded coherence data; S3, selecting a pair of SAR image pairs from the time-space baseline network as a reference unit, and calculating a displacement result by a pixel displacement tracking technology and determining the best matching window parameters of the reference unit; S4, establishing a mapping function between the coherence and the best matching window parameters of the reference unit according to the obtained geocoded coherence data of the reference unit and the determined best matching window parameters of the reference unit; S5, according to the established mapping function, substituting the geocoded coherence data of the remaining SAR image pairs in the time-space baseline network to calculate the best matching window parameters of the remaining SAR image pairs, thereby determining the best matching window parameters of all SAR image pairs in the time-space baseline network; S6, using the determined best matching window parameters of all SAR image pairs, performing pixel displacement tracking processing on all SAR image pairs in the time-space baseline network to obtain the line-of-sight displacement field of each SAR image pair; S7, based on the small baseline set technology, performing time series solution on the obtained line-of-sight displacement field of each SAR image pair to obtain the line-of-sight average deformation rate field of ascending and descending tracks; S8, combining the line-of-sight average deformation rate fields of the ascending and descending tracks, and inversely calculating the surface three-dimensional flow deformation field of the target area by a geometric vector decomposition model.

2. The three-dimensional flow velocity field inversion method based on the coherence-assisted SAR adaptive offset tracking according to claim 1, wherein, The preprocessing comprises: based on a projection equation, using satellite orbit parameters to crop, resample and coordinate system conversion on the digital elevation model data, so that the spatial range and reference system are consistent with the SAR data; selecting one SAR image as a super master image, and precisely registering the remaining multi-scene SAR images to the radar coordinate system of the super master image; based on a preset time baseline threshold and a spatial vertical baseline threshold, combining and screening all registered SAR images in pairs to construct the time-space baseline network.

3. The three-dimensional flow velocity field inversion method based on the coherence-assisted SAR adaptive offset tracking according to claim 1, wherein, The S2 step comprises: S21, for each pair of SAR image pairs in the time-space baseline network, determining the master image and the slave image in time sequence, performing multi-view, registration and interference on the two images to generate a differential interference graph, and removing the terrain phase and flat phase by using the digital elevation model data and the orbit parameters; S22, filtering the generated differential interference graph to obtain a filtered differential interference graph, and calculating the coherence coefficient of each pixel to obtain the coherence data; S23, based on the satellite orbit parameters and the digital elevation model data, converting the obtained filtered differential interference graph and the obtained coherence data from the radar coordinate system to the target geodetic coordinate system to complete geocoding and obtain the geocoded coherence data.

4. The three-dimensional flow velocity field inversion method based on the coherence-assisted SAR adaptive offset tracking according to claim 1, wherein, The best matching window parameters are determined by the pixel displacement tracking technology, which comprises: On the image corresponding to the reference unit, a series of matching windows with different sizes are used to track pixel offset calculation respectively; Compare the displacement field results of each size window; Select the window size corresponding to the best quality evaluation result as the best matching window parameter.

5. The three-dimensional flow velocity field inversion method based on the coherence-assisted SAR adaptive offset tracking according to claim 1, wherein, The mapping function is: ； In the formula, : Best matching window size; : Coherence.

6. The three-dimensional flow velocity field inversion method based on the coherence-assisted SAR adaptive offset tracking according to claim 1, wherein, The following formula is used to calculate the best matching window parameter of the SAR image to be processed in the S5 step: ； In the formula, : Best offset tracking matching window of reference cell; : Coherence of reference cell; : Offset tracking best matching window of SAR image pair to be processed; : Coherence of SAR image pair to be processed.

7. The three-dimensional flow velocity field inversion method based on the coherence-assisted SAR adaptive offset tracking according to claim 1, wherein, The S7 step includes: S71, constructing an observation equation group H·V=δ with the line-of-sight displacement of each SAR image pair as the observation value and the average deformation rate of each adjacent time interval as the unknown quantity, wherein δ is an observation vector composed of all SAR image pair displacement values, V is the average deformation rate sequence vector to be solved, and H is a design matrix; S72, singular value decomposition method is used to solve the observation equation group, and the average line-of-sight deformation rate field of the ascending track and the descending track is obtained respectively.

8. The three-dimensional flow velocity field inversion method based on the coherence-assisted SAR adaptive offset tracking according to claim 1, wherein, The geometric vector decomposition model is a constraint method based on the parallel flow hypothesis.

9. A system for three-dimensional flow velocity field inversion based on coherence-aided SAR adaptive offset tracking, the system being configured to perform the method of three-dimensional flow velocity field inversion based on coherence-aided SAR adaptive offset tracking according to any one of claims 1 to 8, characterized in that It includes: A data preprocessing and network construction module configured to perform the S1 step; A coherence calculation module configured to perform the S2 step; An adaptive window decision module configured to perform the S3, S4 and S5 steps; A displacement field extraction and time series solution module configured to perform the S6 and S7 steps; And, A three-dimensional inversion module configured to perform the S8 step.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to realize the three-dimensional flow velocity field inversion method based on the coherence auxiliary SAR adaptive offset tracking according to any one of claims 1 to 8.

Images