Satellite interferometry method for deformation of frozen soil in season
阅读说明：本技术 冻土季节形变的卫星干涉测量方法 (Satellite interferometry method for deformation of frozen soil in season ) 是由 刘国祥 张瑞 吴婷婷 王晓文 向卫 于 2020-06-10 设计创作，主要内容包括：本发明公开了一种冻土季节形变的卫星干涉测量方法,其包括将冻土区域的SAR影像进行配准并重采样至同一空间坐标系中；将转换后SAR影像进行配对,选取满足设定条件的干涉对进行干涉；采用精密轨道数据移除干涉图中各点的参考椭球面相位,采用高精度DEM模拟地形相位,并移除干涉图中各点的地形相位得到差分干涉图；对差分干涉图进行滤波,并采用最小费用流方法对滤波后的干涉图进行相位解缠,将解缠相位转换为地表形变；获取每个像素在干涉对时间序列中超过预设相干系数的数量,并选取数量大于预设数量的像素点作为短时高相干点；根据选出的短时高相干点,逐点利用Stacking方法基于线性模型求解冻土季节形变的平均速率与累积量。(The invention discloses a satellite interferometry method of frozen soil seasonal deformation, which comprises the steps of registering and resampling SAR images of a frozen soil area to the same space coordinate system; matching the converted SAR images, and selecting an interference pair meeting set conditions for interference; removing the reference ellipsoid phase of each point in the interferogram by adopting the precise orbit data, simulating the terrain phase by adopting a high-precision DEM (digital elevation model), and removing the terrain phase of each point in the interferogram to obtain a differential interferogram; filtering the differential interferogram, performing phase unwrapping on the filtered interferogram by adopting a minimum cost flow method, and converting an unwrapped phase into surface deformation; acquiring the number of pixels exceeding a preset coherence coefficient in an interference pair time sequence, and selecting pixel points with the number larger than the preset number as short-time high coherence points; and according to the selected short-time high coherence points, solving the average rate and cumulant of the seasonal deformation of the frozen soil point by using a Stacking method based on a linear model.)
1. The satellite interferometry method of the seasonal deformation of the frozen soil is characterized by comprising the following steps of:
s1, acquiring a plurality of SAR images in a seasonal deformation frost heaving/thaw sinking stage of the frozen soil region, registering all the SAR images, and resampling to the same space coordinate system;
s2, pairing all SAR images converted into the same space coordinate system in pairs at will to form a plurality of interference pairs, and then selecting the interference pair combination meeting the set conditions for interference to obtain an interference pattern;
s3, removing the reference ellipsoid phase of each point in the interferogram by adopting the precise orbit data, then simulating the terrain phase by adopting a high-precision DEM (digital elevation model), and removing the terrain phase of each point in the interferogram to obtain a differential interferogram;
s4, filtering the differential interferogram by adopting a self-adaptive filtering method, then performing phase unwrapping on the filtered differential interferogram by adopting a minimum cost flow method, and finally converting an unwrapped phase into surface deformation;
s5, acquiring the number of pixels exceeding a preset coherence coefficient in an interference pair time sequence according to the coherence coefficient of each interference image, and selecting pixel points with the number larger than the preset number as short-time high coherence points;
and S6, solving the average rate and cumulant of the seasonal deformation of the frozen soil by using a Stacking method point by point based on a linear model according to the selected short-time high coherence point.
2. The method for satellite interferometry of seasonal deformation of frozen earth according to claim 1, wherein the calculation formula for converting the unwrapping phase into the surface deformation is as follows:
wherein the content of the first and second substances,
3. The satellite interferometry method of seasonal deformation of frozen earth of claim 1, wherein a coherence coefficient of each interferogram is calculated according to a formula:
wherein, M (i, j) and S (i, j) are respectively complex values on pixels i, j on the two SAR images;is complex conjugate multiplication; l is an interference pair serial number; l is the total number of interference pairs; m × n is the window size for calculating the coherence coefficient.
4.The method of claim 2 or 3, wherein the average rate of seasonal deformation of frozen earth is determined
wherein, Δ tkIs the kth interference pair image time interval; q is the number of interference pairs; dkThe ground surface deformation of the k-th interference pair image is obtained.
5. The method for satellite interferometry of seasonal deformation of frozen earth of claim 4, wherein the setting conditions are: interference pairs of the image time base line in a set time and interference pairs with the average coherence coefficient larger than 0.4.
6. The method of claim 1, wherein the resampling of all SAR images after registration to the same spatial coordinate system further comprises:
randomly selecting an SAR image as a main image, registering the rest SAR images with the main image, and resampling the registered SAR images to a space coordinate system where the main image is located.
7. The method for measuring the satellite interferometry of the seasonal deformation of the frozen soil according to claim 4, wherein the high-precision DEM is a Reference 3D high-precision DEM generated by a SPOT-5 French satellite.
8. The method of claim 7, wherein the high accuracy DEM is converted to WGS84 reference before simulating the terrain phase.
Technical Field
The invention relates to the technical field of research on seasonal deformation of frozen soil, in particular to a satellite interferometry method for seasonal deformation of frozen soil.
Background
Along with global warming, global frozen soil is undergoing serious degradation, the problem of frozen soil degradation faced by Qinghai-Tibet plateau is more severe, and plateau frozen soil degradation is mainly manifested by increased average ground temperature, reduced area of permafrost, increased thickness of active layer, increased lower limit of permafrost and the like. Frozen soil degradation causes a series of natural disasters and ecological environmental problems. In the context of global warming, as the frozen soil melts, carbohydrates contained in the frozen soil may undergo a series of chemical reactions and are finally discharged into the atmosphere in the form of carbon substances such as CH4 and CO2, thereby further increasing the global warming and having a significant influence on the global warming.
In addition, the degradation of frozen soil can cause environmental problems such as soil water loss, plateau ecological environment damage, soil organic matter loss, desertification and the like. In addition to this, frozen soil degradation is usually accompanied by various geological disasters, threatening the safety of lives and properties of people and the safety and stability of frozen soil engineering (Qinghai-Tibet railways and Qinghai-Tibet highways). Therefore, the method is particularly urgent for monitoring and researching the permafrost environment of the Qinghai-Tibet plateau, and has very important scientific and practical significance.
The near-surface soil (rock) layer of the frozen soil active layer, which is melted in summer and frozen in winter, is the most dynamic rock layer of the frozen soil region in thermal dynamics. The most direct manifestation of frozen soil freezing and thawing is that the frozen soil is raised and sunk seasonally, the seasonal deformation of the frozen soil is related to the thickness of an active layer of the frozen soil, large-area seasonal deformation of the frozen soil is obtained, and then the inversion of the thickness of the active layer becomes a hot spot of research on the frozen soil. The traditional method for monitoring the deformation of frozen soil mainly comprises the following steps: geodetic methods such as leveling and GPS observation, and methods for burying instruments, geophysical prospecting, and electromagnetic waves. However, the frozen soil environment of the Qinghai-Tibet plateau is severe, the terrain is severe, the air pressure is low, and the air is thin, so that the traditional measurement technology needs to consume a large amount of manpower and financial resources, and large-scale monitoring and research cannot be carried out. Therefore, it is a very challenging problem to perform large space-time scale and high precision surface deformation measurement in frozen soil areas.
With the progress of the remote sensing technology, particularly the development of the synthetic aperture radar technology, a new mode and selection are provided for the research of the frozen soil environment of the Qinghai-Tibet plateau. The synthetic aperture radar Differential interferometry (DInSAR) method has attracted attention in the research of deformation measurement in frozen soil areas in recent years due to its specific advantages, such as high measurement accuracy, low measurement cost, long measurement period, large measurement range, etc. However, in permafrost regions in Qinghai-Tibet plateau, the DInSAR technology is affected by time incoherent, space incoherent and atmospheric delay, and the precision of the DInSAR technology cannot meet the requirement.
In recent decades, time series InSAR analysis methods represented by Permanent Scatterer Interferometry (PSI) and small baseline set (SBAS) methods have been widely developed. The time sequence InSAR technology acquires the ground surface settlement information by detecting the phase characteristics of scattering points of ground surface stable points, can overcome the defects of the traditional DInSAR, and can reach millimeter-level theoretical precision. However, the soil water content in the permafrost region of the Qinghai-Tibet plateau is large, and the soil property in the permafrost region is obviously influenced by the temperature. The earth surface deformation monitoring of the frozen soil region based on the time sequence InSAR is often influenced by phase time decorrelation, and it is difficult to extract a ground target which is stable in radar wave scattering for a long enough time, so that the seasonal deformation resolving precision and reliability of the frozen soil region are influenced.
Disclosure of Invention
Aiming at the defects in the prior art, the satellite interferometry method for the seasonal deformation of the frozen soil provided by the invention solves the problem that the monitoring of the deformation of the earth surface in the frozen soil area is influenced by phase time loss.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that:
a satellite interferometry method for seasonal deformation of frozen earth is provided, which comprises the following steps:
s1, acquiring a plurality of SAR images in a seasonal deformation frost heaving/thaw sinking stage of the frozen soil region, registering all the SAR images, and resampling to the same space coordinate system;
s2, pairing all SAR images converted into the same space coordinate system in pairs at will to form a plurality of interference pairs, and then selecting the interference pair combination meeting the set conditions for interference to obtain an interference pattern;
s3, removing the reference ellipsoid phase of each point in the interferogram by adopting the precise orbit data, then simulating the terrain phase by adopting a high-precision DEM (digital elevation model), and removing the terrain phase of each point in the interferogram to obtain a differential interferogram;
s4, filtering the differential interferogram by adopting a self-adaptive filtering method, then performing phase unwrapping on the filtered differential interferogram by adopting a minimum cost flow method, and finally converting an unwrapped phase into surface deformation;
s5, acquiring the number of pixels exceeding a preset coherence coefficient in an interference pair time sequence according to the coherence coefficient of each interference image, and selecting pixel points with the number larger than the preset number as short-time high coherence points;
and S6, solving the average rate and cumulant of the seasonal deformation of the frozen soil by using a Stacking method point by point based on a linear model according to the selected short-time high coherence point.
The invention has the beneficial effects that: according to the scheme, during seasonal deformation prediction of the frozen soil region, a single seasonal stage (frost heaving/thaw sinking stage) is selected, and then high-precision DEM data is cooperatively used, so that the error control of the scheme can meet a corresponding linear model, the reliability and the precision of an InSAR technology on a surface deformation extraction result of the region can be effectively improved through a mode of high coherence interference pair unwrapping and short-time high coherence point selection, the successful acquisition of seasonal deformation of the frozen soil region is guaranteed, and weather interference is avoided in the prediction process.
Drawings
FIG. 1 is a flow chart of a satellite interferometry method of seasonal deformation of frozen earth.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.
Referring to FIG. 1, FIG. 1 shows a flow chart of a method of satellite interferometry of seasonal variations in frozen earth; as shown in fig. 1, the method S includes steps S1 to S6.
In step S1, acquiring a plurality of SAR images at the seasonal deformation frost heaving/thaw sinking stage of the frozen soil region, registering all the SAR images, and resampling to the same spatial coordinate system; the method specifically comprises the following steps:
randomly selecting an SAR image as a main image, registering the rest SAR images with the main image, and resampling the registered SAR images to a space coordinate system where the main image is located.
In step S2, any pair of SAR images converted into the same spatial coordinate system are paired to form a plurality of interference pairs, and then interference pairs satisfying the set conditions are selected to be combined for interference to obtain an interference pattern.
Experiments compare the same image as a main image, the average coherence coefficient of interference pairs of time baselines of 11 days, 22 days and 33 days in a research area is researched, and the fact that the interferogram of the time baseline of 11 days can ensure higher coherence is found, and unwrapping can be carried out easily; the coherence of the 22-day interferogram is reduced dramatically, and the unwrapping precision is reduced greatly relative to 11 days; a 33 day intervention is not useful for almost complete loss of coherence.
Therefore, the preferable setting conditions of the scheme are as follows: interference pairs with the image time base line within 30 days and interference pairs with the average coherence coefficient larger than 0.4; the best realization mode of the scheme is to select the combination of the 11-day time base line interference pair and the 22-day time base line interference pair with the average coherence coefficient of more than 0.4.
In step S3, the precise orbit data is used to remove the reference ellipsoid phase of each point in the interferogram, and then the high-precision DEM is used to simulate the terrain phase, and the terrain phase of each point in the interferogram is removed to obtain a differential interferogram; wherein the phase of each point in the interferogramComprising the following components:
wherein the content of the first and second substances,
By adopting the implementation mode of the step S3, the difference interference pattern of the elevation residual error can be avoided, so that the deformation phase occupies the dominant component, and the reliability of the deformation rate calculated in the Sacking operation is ensured.
In implementation, the Reference 3D high-precision DEM generated by the SPOT-5 French satellite is preferably selected as the high-precision DEM. The following is a comparative description of the effect of the obtained differential interferogram with Reference 3D high-precision DEM:
during the experiment, firstly, the SRTM DEM is used for carrying out an interference experiment, and when the flat ground effect and the terrain phase are removed, the fact that dense terrain-related fringes still remain in the interference pattern obtained by adopting the method is found, and the vertical precision of the SRTM DEM in a high-altitude area is difficult to meet the requirement.
According to the scheme, after the Reference 3D high-precision DEM is adopted to replace a conventional DEM, experiments are carried out on frozen soil in multiple regions, the error in the elevation of the Reference 3D can reach 1.95-3 m, after the Reference 3D high-precision DEM is used, residual fringes disappear, and the difference interference graph obtained by the selected Reference 3D high-precision DEM can meet the requirement of subsequent Stacking resolving.
It should be noted that Reference 3D DEM elevation Reference is EGM96, and the SAR image in this embodiment is WGS84 Reference, so the Reference 3D high-precision DEM needs to be converted into WGS84 Reference before simulating the terrain phase.
In step S4, the differential interferogram is filtered by an adaptive filtering method, then the filtered differential interferogram is phase unwrapped by a least-cost-stream method, and finally the unwrapped phase is converted into a surface deformation:
wherein the content of the first and second substances,to unwrapping the differential interference phase; d is the deformation of the earth surface; λ is the radar wavelength at which the SAR image is acquired.
In step S5, according to the coherence coefficient of each interferogram, the number of pixels exceeding the preset coherence coefficient in the interference pair time sequence is obtained, and the pixel points whose number is greater than the preset number are selected as short-time high coherence points.
According to the scheme, the interference pairs with the wrong unwrapping are eliminated, the remaining interference pairs are subjected to short-time high coherence point selection, the interference of noise points can be further effectively eliminated through the selection of the high coherence points, the interference pairs easily cause errors in low coherence areas, and the final selection of the coherence points in the areas can be influenced, so that the reliability of results can be effectively improved.
In an embodiment of the present invention, the method for obtaining the short-term high coherence point further includes:
for each pixel, a window of suitable size (e.g., pixel) is selected, and the corresponding coherence factor in each interferogram can be calculated within the window by:
wherein, M (i, j) and S (i, j) are respectively complex values on pixels i, j on the two SAR images;
After obtaining the corresponding coherence coefficient of each pixel in L interference pairs, an appropriate threshold value gamma is setcritFor the value of the coherence coefficient on each pixel time series, the pixel is a short-time high coherence point if the following relationship is satisfied:
C(γl>γcrit)>T,(l＝1,2,…,L)；
wherein C (·) is the counting of variables; t is a set quantity threshold; l is the interference pair number.
The scheme selects the short-time high coherence points by adopting the mode, is different from the traditional coherence coefficient threshold method, the traditional method sets a threshold value for the minimum value of the time sequence coherence coefficient of the pixel, and the scheme judges the number of the threshold values exceeding the specific coherence coefficient in the time sequence, can select the short-time high coherence points and better accords with the characteristic that the interference pattern is less in the scheme.
In step S6, according to the selected short-time high coherence points, solving the average rate and cumulant of seasonal deformation of the frozen soil point by using a Stacking method based on a linear model; solving the average speed of deformation in frozen soil seasonAnd cumulative amountThe linear model of (a) is:
wherein, Δ tkIs the kth interference contrastLike a time interval; q is the number of interference pairs; dkThe ground surface deformation of the k-th interference pair image is obtained.
In the scheme, the seasonal deformation of the frozen soil is calculated, and an SAR image of the frozen soil in the thawing and settling period is preferably selected, namely the frozen soil in the period shows a thawing and settling trend in the freezing and thawing phenomenon, namely the frozen soil in the period has a consistent settling trend. Therefore, in this calculation for a short period of time,
In conclusion, the scheme effectively improves the reliability and the precision of the extraction result of the InSAR technology on the surface deformation of the area in a mode of high coherence interference pair unwrapping and short-time high coherence point selection, and ensures the successful acquisition of seasonal deformation quantity of the frozen soil area.