CN-121981887-A - Ground surface deformation inversion method based on overlap region spliced InSAR deformation field

CN121981887ACN 121981887 ACN121981887 ACN 121981887ACN-121981887-A

Abstract

The invention provides an earth surface deformation inversion method based on an InSAR deformation field spliced by an overlapping area, which comprises the steps of firstly obtaining multi-frame LOS images, image frame information and GNSS observation data in a research area range, classifying the LOS images to obtain an identical track data set, then screening the overlapping area to obtain the identical track overlapping area and corresponding GNSS site data thereof, constructing an observation equation to solve and eliminate errors of the identical track data, splicing the identical track data, carrying out spatial interpolation and projection on the GNSS three-dimensional deformation observation data to obtain GNSS three-dimensional deformation projection data, combining the GNSS three-dimensional deformation projection data, carrying out weighted balancing of an error plane, carrying out adjacent track data splicing, carrying out earth surface three-dimensional deformation inversion to obtain earth surface three-dimensional deformation data and calculating strain rate data to obtain an evaluation result of earth surface three-dimensional deformation inversion. The method improves the accuracy of inversion of the three-dimensional deformation of the earth surface.

Inventors

XU GUANGYU
LIU XUPENG
XIE XIAOWEI
WANG LEYANG

Assignees

东华理工大学南昌校区

Dates

Publication Date: 20260505
Application Date: 20260409

Claims (9)

1. The earth surface deformation inversion method based on the overlap region spliced InSAR deformation field is characterized by comprising the following steps of: step S1, acquiring multi-frame LOS images, corresponding image frame information and GNSS observation data in a research area, wherein the GNSS observation data comprise observation value data, uncertainty of the observation value, GNSS three-dimensional deformation observation data and GNSS site data, and the image frame information comprises coordinates, track types and track numbers of the LOS images; Step S2, classifying the LOS images according to the track types and track numbers, classifying the LOS images with the same track types and track numbers into a group of same-track data to obtain a same-track data set, screening out overlapping areas of two LOS images adjacent to geographic positions in the same-track data according to the same-track data set to obtain overlapping areas of the same track, and obtaining GNSS site data and GNSS three-dimensional deformation observation data in the overlapping areas of the same track; Step S3, constructing an observation equation according to GNSS site data and GNSS three-dimensional deformation observation data in an overlapping area of the same orbit, solving an error plane between two adjacent LOS images in the geographic position in the same orbit data by using the observation equation, eliminating errors between the adjacent LOS images and performing same orbit splicing to obtain strip-shaped LOS images spliced along the orbit direction; Step S4, performing spatial interpolation on the GNSS three-dimensional deformation observation data, and then projecting the spatially-interpolated GNSS three-dimensional deformation observation data onto a corresponding orbit plane where each LOS image is located to obtain GNSS three-dimensional deformation projection data; s5, obtaining strip-shaped LOS images spliced along the track direction corresponding to two adjacent tracks at geographic positions, forming adjacent track data, combining GNSS three-dimensional deformation projection data, and solving an error plane between the two strip-shaped LOS images of the adjacent tracks; S6, constructing an earth surface three-dimensional deformation inversion equation by combining the spliced LOS images covering the research area range and the GNSS three-dimensional deformation observation data after spatial interpolation, and carrying out earth surface three-dimensional deformation inversion to obtain earth surface three-dimensional deformation data; and S7, calculating strain rate data according to the three-dimensional deformation data of the earth surface to obtain an evaluation result of inversion of the three-dimensional deformation of the earth surface.
2. The earth surface deformation inversion method based on the overlap region spliced InSAR deformation field according to claim 1, wherein in the step S1, the GNSS three-dimensional deformation observation data comprise GNSS east-west deformation data, GNSS north-south deformation data and GNSS vertical deformation data, the GNSS site data comprise GNSS site number and GNSS site position information, and the orbit type comprises an ascending orbit type and a descending orbit type.
3. The earth surface deformation inversion method based on the overlap region spliced InSAR deformation field according to claim 2, wherein the step S3 specifically comprises the following steps: Step S31, defining a threshold value of the number of GNSS sites as M GNSS ; Step S32, judging whether the number of GNSS sites in the overlapping area of two LOS images adjacent to each other in the geographic position in the same track data is larger than a threshold M GNSS according to the GNSS site data in the overlapping area of the same track; If the number of GNSS sites in the overlapping area of two LOS images adjacent to each other in the geographic position in the same track data is smaller than a threshold M GNSS , directly constructing an observation equation according to the coordinates of the two LOS images and the observation value data; If the number of GNSS sites in the overlapping area of two LOS images with adjacent geographic positions in the same track data is greater than or equal to a threshold M GNSS , combining the coordinates of the two LOS images, the observation value data and the GNSS three-dimensional deformation observation data to construct an observation equation; solving an error plane between two LOS images adjacent to each other in the geographic position in the same track data by using an observation equation, and eliminating errors between the adjacent LOS images; and step S33, splicing the LOS images of the same track after the errors are eliminated, and obtaining a strip-shaped LOS image spliced along the track direction.
4. The earth surface deformation inversion method based on the overlap region spliced InSAR deformation field according to claim 3, wherein in the step S32, an observation equation directly constructed according to coordinates and observation value data of two LOS images is as follows: , In the formula, Representing an error plane between two LOS images adjacent to each other in geographic positions in the same track data; And Respectively representing the observed value data of two LOS images adjacent to each other in the geographic position in the same track data; And Respectively representing the X-axis coordinate and the Y-axis coordinate of the t pixel point in the overlapped area of the two LOS images; 、 And Is a model factor of an error plane between two LOS images adjacent to each other in the geographic position in the same track data; Then, solving an error plane between two adjacent LOS images in the same track data at the geographic positions by using an observation equation, and eliminating the error between the adjacent LOS images comprises the following specific steps: Solving according to the observation equation to obtain a model factor 、 And Then, the two LOS images are defined as a whole, and then the formula is utilized To calculate observed value data for eliminating an error plane between two LOS images, wherein, The X-axis coordinates representing the entirety of the two LOS images, Representing the Y-axis coordinates of the entirety of the two LOS images, Representing that error planes between two LOS images adjacent to each other in geographic position in co-orbit data are at coordinates The following observed value data; Then, the error planes between two LOS images with adjacent geographic positions in the same track data are distributed averagely, the error between the two LOS images is eliminated, and the observed value data after eliminating the error planes are respectively for the two LOS images And 。
5. The earth surface deformation inversion method based on the overlap region spliced InSAR deformation field of claim 3, wherein in the step S32, an observation equation constructed by combining coordinates of two LOS images, observation value data and GNSS three-dimensional deformation observation data is as follows: , In the formula, And Respectively representing the observed value data of two LOS images adjacent to each other in the geographic position in the same track data; representing GNSS east-west deformation data in the GNSS three-dimensional deformation observation data, Representing GNSS north-south deformation data in the GNSS three-dimensional deformation observation data, Representing GNSS vertical deformation data in the GNSS three-dimensional deformation observation data; And Respectively representing satellite observation azimuth angles corresponding to two LOS images adjacent to geographic positions in the same orbit data; And Respectively representing satellite observation incident angles corresponding to two LOS images adjacent to geographic positions in the same orbit data; And Respectively representing the X-axis coordinate and the Y-axis coordinate of the t pixel point in the overlapped area of the two LOS images; representing the east-west deformation of the t pixel point in the overlapped area of the two LOS images, Representing the north-south deformation of the t pixel point in the overlapped area of the two LOS images, Representing the deformation of the t pixel point in the overlapped area of the two LOS images in the vertical direction; 、、 And 、、 Model factors representing error planes between two LOS images adjacent to geographic positions in the same-track data and a GNSS three-dimensional deformation observation data plane respectively; Then, solving an error plane between two adjacent LOS images in the same track data at the geographic positions by using an observation equation, and eliminating the error between the adjacent LOS images comprises the following specific steps: Solving according to the observation equation to obtain a model factor 、、 And 、、 Then, taking the GNSS three-dimensional deformation observation data plane as a reference plane for eliminating errors, respectively eliminating error planes between two LOS images and the GNSS three-dimensional deformation observation data plane, thereby eliminating the error planes between the two LOS images, wherein the method comprises the following specific steps: Using the formula Sum formula Respectively calculating the observed value data of an error plane between two LOS images and a GNSS three-dimensional deformation observed data plane, wherein, And Respectively representing the X-axis coordinate and the Y-axis coordinate of one of the LOS images, Representing the coordinates of the error plane between the corresponding LOS image and the GNSS three-dimensional deformation observation data plane The following observed value data; And Respectively representing the X-axis and Y-axis coordinates of another LOS image, Representing the coordinates of the error plane between the corresponding LOS image and the GNSS three-dimensional deformation observation data plane The following observed value data; for two LOS images, the observed value data after eliminating the error plane are respectively And 。
6. The earth surface deformation inversion method based on the overlap region spliced InSAR deformation field according to claim 5, wherein the step S4 specifically comprises the following steps: Step S41, screening the GNSS three-dimensional deformation observation data in the research area according to the uncertainty of the observation value, and then performing spatial interpolation to obtain spatially-interpolated GNSS three-dimensional deformation observation data, including spatially-interpolated GNSS east-west deformation data GNSS south-north deformation data after spatial interpolation And spatially interpolated GNSS vertical deformation data ; Step S42, projecting the spatially interpolated GNSS three-dimensional deformation observation data onto the corresponding orbit plane where each LOS image is located, and obtaining GNSS three-dimensional deformation projection data, wherein the projection formula is as follows: , In the formula, Representing the data of the spatially interpolated GNSS three-dimensional deformation observation data projected to the orbit plane where the LOS image is located, namely GNSS three-dimensional deformation projection data; And Respectively corresponding to azimuth angle and incidence angle of observation data during satellite flight; representing spatially interpolated GNSS east-west deformation data, Representing spatially interpolated GNSS north-south deformation data, Representing spatially interpolated GNSS vertical deformation data.
7. The earth surface deformation inversion method based on the overlap region spliced InSAR deformation field according to claim 6, wherein the step S5 specifically comprises the following steps: Step S51, obtaining strip-shaped LOS images spliced along the track direction and corresponding to two adjacent tracks at geographic positions, forming adjacent track data, combining GNSS three-dimensional deformation projection data, respectively solving error planes between the two strip-shaped LOS images of the adjacent tracks in the adjacent track data and the GNSS three-dimensional deformation projection data plane, and determining model factors of error planes between the two strip-shaped LOS images of the adjacent tracks and the GNSS three-dimensional deformation projection data plane, wherein the calculation formula is as follows: , In the formula, Representing the error plane between two elongated LOS images of adjacent tracks, Representing the error plane between one of the elongated LOS images and the GNSS three-dimensional deformed projection data plane, Representing an error plane between another long LOS image and the GNSS three-dimensional deformation projection data plane; And Observation value data of two strip-shaped LOS images respectively representing adjacent tracks; And GNSS three-dimensional deformation projection data on two adjacent tracks in the adjacent track data are respectively represented, and the GNSS three-dimensional deformation projection data of two strip-shaped LOS images of the adjacent tracks are respectively corresponding to the GNSS three-dimensional deformation projection data; And X-axis coordinates and Y-axis coordinates of a t-th pixel point in an overlapping region of two strip-shaped LOS images representing adjacent tracks; 、、 And 、、 Model factors representing error planes between two strip LOS images of adjacent tracks and a GNSS three-dimensional deformation projection data plane respectively; solving to obtain model factors 、、 And 、、 Then, taking the GNSS three-dimensional deformation projection data plane as a reference plane for eliminating errors, respectively eliminating error planes between two strip-shaped LOS images and the GNSS three-dimensional deformation projection data plane, thereby eliminating the error planes between the two strip-shaped LOS images, wherein the method specifically comprises the following steps: Using the formula Sum formula Respectively calculating the observed value data of an error plane between the two strip LOS images and the GNSS three-dimensional deformation projection data plane, wherein, And Respectively representing the X-axis coordinate and the Y-axis coordinate of one of the long LOS images, Representing the coordinates of the error plane between the corresponding long LOS image and the GNSS three-dimensional deformation projection data plane The following observed value data; And Respectively representing the X-axis coordinate and the Y-axis coordinate of another long LOS image, Representing the coordinates of the error plane between the corresponding long LOS image and the GNSS three-dimensional deformation projection data plane The following observed value data; For two long-strip LOS images, after eliminating error planes between the long-strip LOS images and the GNSS three-dimensional deformation projection data plane, the obtained observed value data are respectively And ; And Respectively representing the obtained observation value data of two strip-shaped LOS images after eliminating the error plane between the strip-shaped LOS images and the GNSS three-dimensional deformation projection data plane; Step S52, using the overlapping area data between the two long-strip LOS images of the adjacent tracks, solving an error plane between the two long-strip LOS images of the adjacent tracks due to the track error, and determining a model factor of the error plane between the two long-strip LOS images of the adjacent tracks due to the track error, wherein the calculation formula is as follows: , In the formula, And Observation value data of two strip-shaped LOS images respectively representing adjacent tracks; And GNSS three-dimensional deformation projection data on two tracks adjacent to each other in geographic positions in adjacent track data are respectively represented, and the GNSS three-dimensional deformation projection data correspond to two strip-shaped LOS images; And X-axis coordinates and Y-axis coordinates of a t-th pixel point in an overlapping region of two strip-shaped LOS images of adjacent tracks are represented respectively; 、 And Model factors representing error planes between two elongated LOS images of adjacent tracks due to track errors; solving to obtain model factors of error planes between two long-strip LOS images of adjacent tracks due to track errors 、 And Then, two long-strip LOS images are defined as a whole, and then the formula is utilized To calculate observed value data for an error plane between two elongated LOS images of adjacent tracks due to track errors, wherein, The X-axis coordinates of the entirety of two elongated LOS images representing adjacent tracks, The Y-axis coordinates of the entirety of two elongated LOS images representing adjacent tracks, Error plane between two long LOS images representing adjacent tracks is at coordinates The lower observation value data is used for eliminating the observation value data of an error plane caused by the track error between two strip LOS images of adjacent tracks; Respectively solving error planes caused by track errors between each long-strip-shaped LOS image and two long-strip-shaped LOS images on the two sides of the track, determining model factors of the error planes caused by the track errors between each long-strip-shaped LOS image and the two long-strip-shaped LOS images on the two sides of the track, and respectively calculating to obtain observed value data for eliminating the error planes caused by the track errors between the long-strip-shaped LOS images and the two long-strip-shaped LOS images on the two sides of the track; the western correction value of the long LOS image is defined as the observed value data for eliminating the error plane caused by the orbit error between the long LOS image and the long LOS image on the western side of the orbit Defining the observed value data for eliminating the error plane between the long LOS image and the long LOS image on the eastern side of the track as the eastern side correction value of the long LOS image, and recording as ; And then, carrying out the weighted balancing of an error plane on each long LOS image according to the size of the area difference of the overlapping areas on the east and west sides of each long LOS image track, wherein the specific steps are as follows: Defining the area of the eastern overlapping region of the long LOS image as S East (Dong) , and the area of the western overlapping region of the long LOS image as S Western medicine ; if the area difference of overlapping areas on the two sides of the long LOS image is S East (Dong) /S Western medicine is less than or equal to 0.8 or S East (Dong) /S Western medicine is less than or equal to 1.2, error plane correction on the two sides is not performed; If the area difference of overlapping areas on the two sides of the long LOS images is 0.8< S East (Dong) /S Western medicine <1.2, defining the correction value of the weighted balancing of the error plane of each long LOS image as The corrected observed value data of the long LOS image is , wherein, Observed value data of a long LOS image after eliminating an error plane caused by an overlapping region; And step S53, carrying out weighted balancing on the error plane of each long-strip LOS image to obtain corrected long-strip LOS images, and carrying out adjacent track splicing on the corrected long-strip LOS images according to the track type to obtain spliced LOS images covering the research area range, wherein the spliced LOS images comprise spliced lifting LOS images covering the research area range and spliced lowering LOS images covering the research area range.
8. The earth surface deformation inversion method based on the overlap region stitching InSAR deformation field according to claim 7, wherein in the step S6, the earth surface three-dimensional deformation inversion equation is as follows: , In the formula, Representing spliced up-track LOS images covering the area of investigation, Represents the spliced derailment LOS image covering the investigation region, Representing the GNSS north-south deformation data after spatial interpolation; The satellite azimuth angle representing the orbit up type data, The satellite azimuth representing the down-track type data, The satellite incidence angle representing the orbit type data, Satellite incidence angle representing the down-track type data; 、 And Representing a three-dimensional deformation rate field of the earth's surface that requires inversion, wherein, Representing the vertical deformation rate field, Representing the north-south deformation rate field, Representing an east-west deformation rate field, which is a value to be solved by a three-dimensional deformation inversion equation of the earth surface, wherein the east-west deformation rate field And a vertical deformation rate field According to spliced track lifting LOS image covering the range of the research area And derailment LOS image Solving to obtain a north-south deformation rate field Based on the spatially interpolated GNSS north-south deformation data And the three groups of deformation rate field data are obtained through solving to jointly form the three-dimensional deformation data of the earth surface.
9. The earth surface deformation inversion method based on the overlap region spliced InSAR deformation field according to claim 8, wherein the step S7 specifically comprises the following steps: step S71, calculating a horizontal strain rate tensor according to the three-dimensional deformation data of the earth surface obtained in the step S6, wherein the calculation formula is as follows: , In the formula, Representing the horizontal strain rate tensor, , , , Representing a velocity gradient in the horizontal direction, wherein, Representing the change rate of the east-west speed along with the change of the X-axis coordinate; representing the change rate of the speed of the north-south direction along with the change of the Y-axis coordinate; = , 、 The two are the same in meaning, are the average effects of measuring the change of the speed in the east-west direction and the change of the speed in the north-south direction, reflecting the shearing deformation of an object, wherein x represents the distance between two adjacent pixels in the middle east-west direction in the horizontal direction, and y represents the distance between two adjacent pixels in the north-south direction in the horizontal direction; indicating that the deformation speed in the east-west direction is deviated from the X-axis direction, Indicating that the deformation speed in the east-west direction is deflected in the Y-axis direction, Indicating that the deformation speed in the north-south direction is deviated from the X-axis direction, Representing the deviation of the deformation speed in the north-south direction to the Y-axis direction; step S72, calculating a second order invariant of the horizontal expansion rate, the shear rate and the horizontal strain rate tensor according to the velocity gradient in each horizontal direction in the horizontal strain rate tensor obtained in step S71, wherein the calculation formula is as follows: , In the formula, Representing the rate of expansion in the horizontal direction, Representing the shear rate of the shear-flow, A second order invariant representing a horizontal strain rate tensor; The second-order invariant data of the horizontal expansion rate, the shear rate and the horizontal strain rate tensor can be used for crust deformation analysis and geological engineering evaluation, wherein the larger the positive value of the horizontal expansion rate is, the stronger the stretching effect is, the obvious surface expansion trend corresponds to a positive fault zone and a crust expansion zone, the larger the negative value of the horizontal expansion rate is, the stronger the extrusion effect is, the obvious surface shortening trend corresponds to a thrust fault zone and a plate extrusion zone, the larger the shear rate value is, the stronger the shearing effect is, the stronger the relative slippage of a crust corresponds to a slip fault zone and a crust boundary slip zone, and the larger the second-order invariant value of the horizontal strain rate tensor is, the larger the comprehensive deformation strength is, the more severe the crust movement corresponds to a fault activity zone and a high-risk seismic zone.

Description

Ground surface deformation inversion method based on overlap region spliced InSAR deformation field Technical Field The invention relates to the technical field of geophysics and remote sensing image processing, in particular to a ground surface deformation inversion method based on an overlapping region spliced InSAR deformation field. Background Geological disasters are complex processes under the combined action of earth dynamics and environmental factors, and involve various deformation events including but not limited to earthquakes, volcanic activities, landslides, ground subsidence, debris flows, ground subsidence, freeze thawing deformation caused by cold and hot season circulation, and the like. Regardless of the disaster type, its development is often accompanied by long-term or short-term strain accumulation, energy release and stress redistribution, which processes exhibit significant non-uniformity and multiscale coupling on a temporal and spatial scale, with different mechanisms of action also interacting through stress transmission, topographical deformations and subsurface medium changes. In order to realize effective risk assessment, early warning response and post-disaster recovery planning, continuous and high-resolution reconstruction and monitoring of the three-dimensional deformation of the earth surface are required in a wide range. The reconstruction can reveal the time-space evolution characteristics of different disaster processes, help infer the geometric shapes of potential faults and landslide zones, the mechanical behaviors of rock and soil bodies and underground fluid permeation paths, explore driving mechanisms and finally support the scenerization hazard simulation and engineering decision. The interference synthetic aperture radar (Interferometric Synthetic Aperture Radar, inSAR) observation technology has become a key means for monitoring the surface deformation due to the advantages of high spatial resolution, wide coverage and relative insensitivity to meteorological conditions. Conventional InSAR observation can only provide deformation information in a radar Line of sight (LOS) direction, and to obtain a three-dimensional deformation field, multi-orbit and multi-view observation data are usually combined, or other observation means such as a global navigation satellite system (Global Navigation SATELLITE SYSTEM, GNSS for short), pixel offset tracking (Pixel offset tracking, POT for short), multi-aperture radar interference (Multiple Aperture Interferometry, MAI for short) or a ground surface deformation model are introduced to carry out constraint solving. However, in large-area, long-time scale deformation inversion, the problem becomes more complex. To obtain a wide range of three-dimensional deformation information, it is necessary to integrate the observation data from a large number of different geographical areas, different orbits and different time windows. Since the geometric relationships of different observation frames are not completely consistent, unstable step phenomenon often occurs in the overlapped area, for example, the observed value of the overlapped area of the same orbit observation frame may deviate, and the overlapped area of different orbit observation frames may have irregular step. These steps can directly affect the subsequent three-dimensional inversion results, reduce the stability and accuracy of the inversion, and affect its time-space consistency. The key point of the large-area earth surface deformation inversion by utilizing InSAR observation data is to splice different observation frames into a complete data set. The core of the splicing operation is to unify the observation frames of different observation frames under a fixed reference system, thereby reducing geometric projection errors caused by frame differences and inhibiting the influence of long wavelengths. At present, aiming at the inversion of large-area ground surface three-dimensional deformation, two strategies are mainly adopted, namely (1) before deformation of each radar sight direction (Line of sight, LOS direction for short) is calculated, a long track window is used for processing an original observation image. This gives a series of successive observation bands, each representing LOS direction deformation information for a track. And then, respectively processing the data of the ascending orbit and the descending orbit, and solving the three-dimensional deformation of the earth surface of the large area by utilizing the data of the ascending orbit and the descending orbit in the overlapped area and combining the incident angle and the azimuth angle of the satellite. (2) And splicing the original observation frames of the multi-frame LOS direction deformation subjected to the standardization processing in the research area, so that the whole research area is covered. And then, splicing LOS direction deformation images according to the data types of the ascending tra