CN-117075107-B - Frozen soil area InSAR deformation monitoring method based on space ICA separation

CN117075107BCN 117075107 BCN117075107 BCN 117075107BCN-117075107-B

Abstract

The invention provides a frozen soil area InSAR deformation monitoring method based on spatial ICA separation, which comprises the steps of firstly obtaining a differential interference phase map after unwrapping of a monitored frozen soil area by utilizing a time sequence InSAR technology, secondly carrying out spatial phase separation of the frozen soil area by utilizing spatial ICA, then respectively modeling the separated multi-year frozen soil area and seasonal frozen soil area by adopting different models, further solving frozen soil deformation parameters by adopting a parameter estimation method based on Jacobian iteration, and finally adding low-pass deformation fitted by the models and filtered high-pass deformation to obtain a deformation monitoring result of the frozen soil area. According to the invention, the ICA technology is used for extracting the phase signals related to the deformation of the frozen soil region in advance for modeling, so that the modeling precision and the deformation estimation precision of the InSAR technology can be improved, and references are provided for comprehensive environment management and disaster assessment of the frozen soil region.

Inventors

XING XUEMIN
Ge Jiawang
Lei Minchao
XING XUEHUI
Shi Jiancun
CAI JIEHUA
ZHENG GUANFENG
Long Jingjian
LI HAOXIAN

Assignees

长沙理工大学

Dates

Publication Date: 20260508
Application Date: 20230810

Claims (7)

1. The method for monitoring InSAR deformation in the frozen soil area based on space ICA separation is characterized by comprising the following steps: Step one, generating a time sequence InSAR interference phase based on a high coherence point, which comprises the steps of obtaining a differential interference phase diagram after monitoring unwrapping of a frozen soil area by utilizing a time sequence InSAR technology; step two, spatial ICA phase separation of the frozen soil area, namely, performing spatial ICA phase separation of the frozen soil area by utilizing the spatial ICA; Modeling the deformation time sequence of InSAR frozen soil, namely respectively modeling the separated multi-year frozen soil area and the seasonal frozen soil area by adopting different models; Step four, estimating deformation parameters of the frozen soil, namely solving the deformation parameters of the frozen soil by using a jacobian iterative parameter estimation method; and fifthly, generating time sequence deformation of the frozen soil region, namely adding the low-pass deformation fitted by the model and the filtered high-pass deformation to obtain a deformation monitoring result of the frozen soil region.
2. The method for monitoring the deformation of the InSAR in the frozen soil area according to claim 1, wherein the first step comprises the following steps: step 1.1, pair Interference combination and baseline estimation are carried out on the scene SAR image data to generate a connection diagram, and then multi-view, image registration and resampling are carried out to generate An interferogram and a coherence coefficient map, wherein: ; Step 1.2, carrying out flattening, topographic removal and filtering treatment on the interference pattern in the step 1.1 to generate a differential interference pattern; step 1.3, carrying out phase unwrapping on the differential interference pattern in the step 1.2 to obtain an unwrapped differential interference pattern; step 1.4, performing high-coherence point extraction by using the coherence map generated in step 1.1 and the unwrapped differential interference map obtained in step 1.3 to obtain And generating a time sequence phase matrix taking the first scene image as a reference based on SVD decomposition, namely monitoring a differential interference phase map after disentangled in the frozen soil region.
3. The method for monitoring the deformation of the InSAR in the frozen soil area according to claim 2, wherein the second step comprises the following steps: Constructing a relation between ICA and InSAR interference phases, wherein the ICA model is expressed as: 1); Wherein: is a mixed signal, namely a time sequence phase matrix taking a first scene image as a reference, and time sequence phase Comprises Column phase, in which The time-series phase is expressed as a matrix of expression 2) for subtracting one from the number of images : 2); In the middle of Is the first Differential phase of the scene image relative to the first scene image; for a hybrid matrix, each row represents the spatial response of an independent component; For an independent component matrix, i.e., an independent component recovered by ICA, each row represents an independent component; Independent component matrix Viewed as a linear combination of mutually independent components, the phases of all high coherence points over the entire time sequence are represented by the individual components: 3); In the middle of For different independent components, in the InSAR data, Comprising independent signal deformation related components Terrain-related component Track related component Atmospheric related component Noise related component Deformation related component And also contains linear deformation component related to earth surface long-term displacement Environmental dependent periodic deformation component As shown in expression 4): 4); according to expression 3), expression 1) is expressed as expression 5): 5); Wherein: for the timing phase referenced to the first scene image, The number of points is high coherence; is a mixing matrix; For the number of separated components; for all independent components obtained by ICA separation, the components are separated from each other Composition; For InSAR signals, the deformation phase and the noise related phase are considered to be independent in time and space, and according to the characteristics of InSAR interference phase components, the InSAR interference phase is decomposed by using FastICA, and the required signals are extracted.
4. The method for monitoring the deformation of the InSAR in the frozen soil area according to claim 3, wherein the method comprises the following steps: Modeling and fitting deformation by adopting a multi-rate linear model aiming at a permafrost region: 6); Wherein: is the permafrost region The deformation of a point of high coherence, For the rate of deformation between adjacent images, For the time interval between adjacent images, ; Fitting deformation of a frozen soil deformation model considering environmental factors is adopted for seasonal frozen soil areas: 7); Wherein: is the seasonal frozen soil area Deformation of the high coherence points; As a function of the temperature parameter(s), The parameters are precipitation parameters, and are obtained through temperature and precipitation data calculation; Respectively is Coefficients.
5. The method for monitoring InSAR deformation in frozen soil area according to claim 4, wherein in step four, for multi-rate linear model, an equation can be built for each high coherence point on the interferogram, according to the time matrix of interference pair Solving rate using jacobian iterative parameter estimation algorithm The method comprises the following steps: step 4.1.1 for a time matrix Proceeding with And (3) decomposition: 8); Wherein: is a diagonal array, and is directly inverted to obtain ; Is a lower triangular matrix; Is an upper triangular matrix; expression 6) is written as: ; Step 4.1.2, further determining the iterative format as expression 9) according to the result obtained in step 4.1.1): 9); Step 4.1.3, according to expression 8), let , Then the expression 9) is rewritten and initialized as follows: 10); the termination conditions of step 4.1.4, expression 10) are , The size of (2) is set according to the actual operation requirement, obtained at this time Is the optimal solution of unknown parameters.
6. The method for monitoring deformation of InSAR in frozen soil area according to claim 4, wherein in the fourth step, the time matrix of interference pairs is known for the deformation model taking environmental factors into consideration And an environmental factor comprising temperature and precipitation, calculating parameters 、 : 11); 12); Wherein: Is the first The temperature parameter of the individual interference pair, Temperature data of master and slave image time on the interference pair respectively; Is the first The precipitation parameters of the individual interference pairs, Precipitation data of master-slave images on the interference pair respectively; Will be Writing into parameter matrix : 13); The unknown parameters are expressed as: solving by using Jacobi iterative parameter estimation algorithm Comprising the following steps: step 4.2.1 for coefficient matrix Proceeding with And (3) decomposition: 14); expression 7) is written as: ; Step 4.2.2, determining an iteration format according to the result of step 4.2.1 as follows: 15); Step 4.2.3, according to expression 14), let , Then the expression 15) is rewritten and initialized for iteration: 16); the termination conditions for step 4.2.4, expression 16) are , The size of (2) is set according to the actual operation requirement, obtained at this time Is the optimal solution of unknown parameters.
7. The method for monitoring the deformation of the InSAR in the frozen soil area according to claim 4, wherein the fifth step is specifically as follows: Step 5.1, fitting the multi-rate linear model obtained in the step three to form deformation Model fitting deformation taking into account environmental factors Transformation phase to multirate linear model fit Deformation phase fitted to model taking into account environmental factors Calculating the residual phase : 17); Wherein: Is the original total interference phase; step 5.2, filtering the residual phase in expression 17) using mean Performing spatial low-pass filtering, triangular filtering and temporal high-pass filtering to obtain high-pass deformation phase, and converting into high-pass deformation phase ; Step 5.3, adding the low-pass and high-pass deformations to obtain the total deformation of the region : 18); Wherein the low-pass deformation The deformation of the frozen soil deformation model fitting comprising a multi-rate linear model and taking environmental factors into account: ; And 5.4, performing geocoding on the coherent point view line to the total deformation quantity, and generating the vertical time sequence deformation of the earth surface of the frozen soil area, thus obtaining the deformation monitoring result of the frozen soil area.

Description

Frozen soil area InSAR deformation monitoring method based on space ICA separation Technical Field The invention relates to frozen earth surface deformation monitoring, in particular to a frozen earth area InSAR deformation monitoring method based on spatial ICA separation. Background The traditional method for monitoring the earth surface subsidence of the frozen soil region, such as a total station/prism method, a leveling method, a global navigation satellite system and the like, can meet the monitoring precision requirement of the frozen soil region. However, the methods have high cost and lower space-time resolution, and have the characteristics that the whole surface subsidence observation is insufficient for frozen soil areas with a slice, and the frequent in-and-out site survey of monitoring personnel is required, so that the application of the method is severely restricted. The differential synthetic aperture radar interferometry (DIFFERENTIAL INTERFEROMETRIC SYNTHETIC APERTURE RADAR, D-InSAR for short) technology is used as an emerging space earth observation technology, is mainly used for monitoring the surface deformation of the earth in the centimeter level or smaller in the radar sight line direction, has the advantages of wide monitoring range, high spatial resolution, non-contact measurement and the like, greatly overcomes the defects of the traditional measurement means, and provides a new opportunity for the settlement monitoring of a large-area frozen soil area. However, the monitoring accuracy of the D-InSAR technique is limited by the effects of space-time uncorrelation and atmospheric delays. In order to overcome the defects of the D-InSAR technology, the time sequence InSAR technology is proposed by scholars, and is mainly represented by a permanent scatterer (PERMANENT SCATTERER, PS for short), a small baseline set (Small Baseline Subset, SBAS for short), time domain related point interference (Temporarily Coherent Point InSAR, TCP-InSAR for short) and other technologies, and by carrying out phase analysis on a plurality of SAR images on the time sequence, the surface deformation extraction is realized by utilizing high quality points with stable scattering characteristics, so that the technology has more potential in monitoring the surface deformation of a frozen soil area. Frozen soil settlement is a complex, nonlinear and obvious periodic process, and the surface deformation presented by the frozen soil settlement is easily influenced by external climatic environments. In the conventional InSAR frozen soil deformation estimation method, an assumption model is mostly utilized to perform equal weight modeling on each deformation component in the interference phase after unwrapping, the frozen soil deformation related phase components are not reasonably extracted according to different physical causes, and the estimation progress is not high. Therefore, the design of the method which has high estimation precision and can be suitable for monitoring the settlement of the frozen soil area has important significance. Disclosure of Invention The invention provides a frozen soil area InSAR deformation monitoring method based on space ICA separation, which aims to overcome the defects of low deformation estimation precision and difficult physical interpretation caused by unreasonable extraction of deformation related components of different frozen soil types in the existing InSAR frozen soil deformation monitoring, and specifically comprises the steps of firstly, obtaining a differential interference phase map after unwrapping of a monitored frozen soil area by utilizing a time sequence InSAR technology; the method comprises the steps of carrying out space phase separation on a frozen soil region by utilizing a space ICA, modeling the separated multi-year frozen soil region and seasonal frozen soil region by adopting different models respectively, solving frozen soil deformation parameters by a parameter estimation method based on Jacobian iteration, and finally, adding low-pass deformation fitted by the models and filtered high-pass deformation to obtain a deformation monitoring result of the frozen soil region. The method for filtering the frozen soil by using time and space can reasonably remove errors such as atmospheric noise, phase noise and the like, effectively extract boundary and time sequence deformation phase signals of different types of frozen soil, improve deformation modeling and deformation estimation precision, is more beneficial to scientific interpretation of deformation, and provides references for comprehensive environment management and disaster assessment of frozen soil areas. The specific scheme is as follows: A frozen soil area InSAR deformation monitoring method based on space ICA separation comprises the following steps: Step one, generating a time sequence InSAR interference phase based on a high coherence point, which comprises the steps of obtaining a differen