Long-term spatiotemporal deformation monitoring method of landslides in reservoir areas based on multi-platform SAR satellites
Through the comprehensive utilization of multi-platform SAR satellite data, the accuracy and direction limitations of long-term spatial deformation monitoring of landslides in the reservoir area are solved, the three-dimensional deformation rate inversion and sliding direction determination of landslides are realized, and the long-term spatial deformation evolution of landslides is restored. Landslide disaster prevention and control provides more precise technical support.
Patent Information
- Application Number
- CN202411311735.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-20
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-09-20
AI Technical Summary
The prior art is difficult to effectively monitor the long-term spatial and temporal deformation of landslides in reservoir areas, especially in vegetation coverage areas. Traditional methods have problems such as insufficient accuracy, limited direction and short monitoring periods, making it difficult to restore the three-dimensional deformation evolution of landslides.
Using multi-platform SAR satellite data, the line of sight and orientation deformation inversion of ALOS-1, Sentinel-1 and TerraSAR-X data is combined with data from different platforms to establish observation equations, reconstruct the average three-dimensional deformation rate of landslides, and calculate the landslide image The sliding direction of the element and the long-term spatial and temporal deformation evolution of the landslide is restored through displacement time series decomposition.
The accuracy and dimension of landslide deformation monitoring is improved, the irrelevant impact of vegetation coverage is reduced, the problem of inconsistent data revisiting cycles on different platforms is solved, and the accurate monitoring and inversion of long-term spatial and temporal deformation of landslides is achieved.
Smart Images

Figure CN118837885B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of landslide deformation monitoring, and in particular to a long-term spatiotemporal deformation monitoring method of reservoir area landslides based on multi-platform SAR satellites. Background Art
[0002] Deformation monitoring is the most direct and effective way to prevent and warn landslide disasters, and is also an important means to understand its evolution mechanism. In the existing landslide deformation monitoring technology, traditional ground measurement methods, such as global navigation satellite systems or leveling, can provide high-precision deformation information of surface monitoring points, but are limited by the cost of manpower and monitoring equipment, and the sparsely distributed discrete monitoring points also limit our comprehensive understanding of the global deformation and evolution mechanism of landslides. In contrast, the booming remote sensing technology in recent years has completely overturned our ability to monitor ground deformation on a global scale, enabling us to obtain high-precision, high-temporal and high-spatial resolution large-scale landslide surface deformation information, which is mainly achieved through synthetic aperture radar interferometry technology.
[0003] However, the current InSAR technology can only measure the projection component of the actual surface deformation along the radar line of sight, and is extremely insensitive to north-south deformation. The sensitivity to the Sentinel-1 orbit-raising data is only 11%. To obtain the true three-dimensional surface deformation, at least three independent observations from two satellite orbits need to be fused, for example, three independent observations including two line-of-sight deformations and one azimuth deformation, or fewer independent observations need to be fused under the premise of a priori kinematic assumptions, such as the assumption of parallel surface motion. Nevertheless, in the case of landslides in reservoir areas that are often covered by seasonal vegetation, restoring the long-term three-dimensional spatiotemporal deformation of landslides is still a huge challenge, mainly due to the following difficulties: (1) InSAR methods are prone to rapid decoherence in vegetated areas, making it difficult to obtain effective measurements; (2) The azimuth monitoring results obtained by multi-aperture InSAR or pixel offset tracking methods are only accurate at the decimeter level, making it difficult to accurately quantify the deformation of slow-moving landslides; (3) The revisit periods of radar satellites on different platforms are inconsistent, making it difficult to obtain multiple independent observation results from different SAR satellites at the same time, making it difficult to achieve time-series inversion of the three-dimensional deformation of the landslide; and (4) There are often large observation gaps between different SAR satellites, making it difficult to restore the long-term deformation evolution of the landslide.
[0004] As can be seen from the above, the existing traditional monitoring methods and remote sensing technologies lack methods that can restore the long-term spatiotemporal deformation evolution of reservoir landslides, which limits further landslide prevention and early warning work. The existing remote sensing monitoring technology does not solve the problems of insufficient accuracy, limited direction, and short monitoring period for reservoir landslide deformation monitoring.
[0005] For example, the patent with publication number CN117872350A discloses a landslide deformation monitoring method, which includes the following steps: 1) a three-dimensional laser scanner obtains the deformation of the monitoring point; 2) the deformation result of the monitoring point obtained by using SBAS-InSAR is converted from the radar coordinate system to the geographic coordinate system; 3) the three-dimensional laser scanning coordinates and the SBAS-InSAR coordinates are converted to the reference coordinate system by using the following formula; 4) the projection angles α, β are obtained; 5) the three-dimensional displacement field is calculated using the obtained projection angles α, β and the high-precision line-of-sight displacement LOS obtained by SBAS-InSAR: the three-dimensional deformation information of the deformable body is obtained by three-dimensional laser scanning, and the high-precision displacement in the radar line of sight is obtained by combining SBAS-InSAR technology, so as to realize the rapid acquisition of high-precision three-dimensional displacement field.
[0006] For example, a Chinese patent with the authorization announcement number CN111474544B discloses a landslide deformation monitoring and early warning method based on SAR data. The method includes: obtaining SAR images of the covered area; selecting common master images; SAR image registration; interferometric pair matching; generating differential interferograms; screening interferometric pairs; removing flat ground phase and terrain phase; removing atmospheric phase and residuals; establishing a model; inverting the sedimentation rate and time series deformation. In view of the fact that the water vapor content in the air will affect the propagation path when the SAR satellite acquires data, thereby causing large errors, the rainfall data covering the experimental area is added to the error image factor in a certain proportion, and a model is established to invert the deformation rate of the SAR data using the least squares method, and the time series deformation of the experimental area within the research period is obtained, thereby reducing the monitoring error caused by factors such as heavy rain, improving the accuracy of InSAR technology in acquiring surface deformation, and providing data and decision support for monitoring and early warning of landslide deformation.
[0007] The above patents have the problems raised by this background technology: the above landslide deformation monitoring method obtains the deformation of the monitoring point through a three-dimensional laser scanner, and converts the deformation results of the monitoring point obtained by SBAS-InSAR from the radar coordinate system to the geographic coordinate system, and converts the three-dimensional laser scanning coordinates and SBAS-InSAR coordinates to the reference coordinate system, thereby calculating the three-dimensional displacement field and realizing the rapid acquisition of high-precision three-dimensional displacement field; the above landslide deformation monitoring and early warning method obtains the SAR image of the covered area, and adds the rainfall data to the error image factor according to a certain proportion, and establishes a model to invert the deformation rate of the SAR data using the least squares method, and obtains the time series deformation of the experimental area during the study period. The above two patents do not solve the problems of insufficient accuracy, directional limitations, and short monitoring period in the existing remote sensing monitoring technology for reservoir landslide deformation monitoring. To solve this problem, the present invention proposes a long-term spatiotemporal deformation monitoring method for reservoir landslides based on multi-platform SAR satellites. Summary of the invention
[0008] In view of the above problems existing in the prior art, the present invention proposes a long-term spatiotemporal deformation monitoring method for reservoir landslides based on multi-platform SAR satellites.
[0009] Therefore, the object of the present invention is to provide a long-term spatiotemporal deformation monitoring method for reservoir landslides based on multi-platform SAR satellites.
[0010] In order to solve the above technical problems, the present invention provides the following technical solutions: collecting multi-platform and multi-band SAR satellite data covering the landslide, and performing landslide deformation inversion according to the wavelength and spatial resolution characteristics of the SAR satellite data, the landslide deformation inversion includes line-of-sight time series deformation inversion of ALOS-1 data and Sentinel-1 data and azimuth deformation rate inversion of TerraSAR-X data, and generating line-of-sight deformation rate and line-of-sight displacement time series according to the line-of-sight time series deformation inversion, and generating azimuth deformation rate according to the azimuth deformation rate inversion;
[0011] The azimuth deformation rate and the sightline deformation rate in the overlapping monitoring interval are combined to establish the observation equation and reconstruct the average three-dimensional deformation rate of the landslide.
[0012] A velocity vector equation is established by reconstructing the average three-dimensional deformation rate, the average velocity vector of each landslide pixel in the geographic coordinate system is calculated, and the sliding direction of each landslide pixel in the monitoring interval is obtained;
[0013] Projecting the sight-direction displacement time series of each landslide pixel to the sliding direction within the monitoring interval one by one, and obtaining the displacement time series of the landslide along the sliding direction;
[0014] The displacement time series of the characteristic points of the landslide along the sliding direction is decomposed into displacement trend terms and displacement period terms to restore the long-term spatiotemporal deformation evolution of the landslide.
[0015] As a preferred solution of the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellites described in the present invention, the SAR satellite data includes data covering different orbital directions of the same landslide and data in a single orbital direction, and maintains two or more independent observation results.
[0016] As a preferred solution of the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellites of the present invention, the azimuthal deformation rate inversion of the TerraSAR-X data includes: performing azimuthal deformation rate inversion on the TerraSAR-X data by combining pixel offset tracking and stacking processing to generate the azimuthal deformation rate, wherein the functional expression of the stacking processing is as follows:
[0017] ;
[0018] In the formula, represents the azimuth average deformation rate of TerraSAR-X data, Indicates The pixel offset tracking inversion results of the offset pairs along the azimuth direction, Indicates The time span of the offset pairs, Indicates the number of offset pairs;
[0019] In the stacking process, the accumulation of displacement terms obtained by tracking and inverting the pixel offset of the offset pair along the azimuth direction changes linearly and is proportional to the number of offset pairs. , while the inverted background noise increases only with the square root of the number of offset pairs and is proportional to , the functional expression of the inversion error of the azimuthal average deformation rate is as follows:
[0020] ;
[0021] In the formula, represents the inversion error of the average deformation rate in the azimuth direction, Indicates The pixel offsets of each offset pair along the azimuth direction track the error in the inversion result.
[0022] As a preferred solution of the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellites of the present invention, the establishment rules of the observation equation include:
[0023] If there are three or more independent observations, an observation equation is established;
[0024] If there are two independent observation results, the observation equation is established in combination with the parallel motion of the surface.
[0025] As a preferred solution of the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellites described in the present invention, there are three or more independent observation results, and the established observation equation is as follows:
[0026] ;
[0027] In the formula, represents the line-of-sight deformation rate of the ascending orbit, represents the line-of-sight deformation rate of the descending orbit, represents the azimuthal deformation rate of the ascending orbit, represents the azimuthal deformation rate of the descending orbit, represents the incident angle of the ascending orbit, represents the incident angle of the descending orbit, Indicates the satellite heading angle of orbit raising, represents the satellite heading angle of descending orbit, represents the inversion error of the observation equation;
[0028] The functional expression of the observation equation established in combination with the parallel motion of the earth's surface is as follows:
[0029] ;
[0030] In the formula, Indicates the line-of-sight deformation rate of Sentinel-1 data, represents the azimuthal deformation rate of TerraSAR-X data, represents the sine function, represents the cosine function, represents the incident angle of Sentinel-1 data, Represents the satellite heading angle of Sentinel-1 data, represents the satellite heading angle of TerraSAR-X data, represents the east-west velocity, represents the north-south velocity, represents the vertical velocity, represents the landslide elevation obtained from the digital elevation model, represents the first-order derivative of the landslide elevation along the north-south direction, Represents the first-order derivative of the landslide elevation along the east-west direction.
[0031] As a preferred solution of the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellite of the present invention, the function expression of the velocity vector equation is as follows:
[0032] ;
[0033] In the formula, represents the unit velocity vector of the landslide pixel, Represents the velocity vector of the landslide pixel in the geographic coordinate system, Represents the L2 norm of the three-dimensional rate.
[0034] As a preferred solution of the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellite of the present invention, the functional expression of the displacement time series of the landslide along the sliding direction is as follows:
[0035] ;
[0036] In the formula, represents the displacement time series of the landslide along the sliding direction, represents the time series of line-of-sight displacement of ALOS-1 data and Sentinel-1 data calculated by the SBAS algorithm, The unit vector representing the radar line of sight direction;
[0037] Among them, the expression of the time series of line of sight displacement is: , Indicates The time series of line-of-sight displacements of the interference pairs is obtained by inverting the line-of-sight time series deformation.
[0038] As a preferred solution of the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellites of the present invention, wherein: the displacement time series of the characteristic points of the landslide along the sliding direction is decomposed into a displacement trend term and a displacement period term, and the long-term spatiotemporal deformation evolution of the landslide is restored, including: decomposing the displacement time series of ALOS-1 data and Sentinel-1 data along the sliding direction respectively by an empirical mode decomposition algorithm, and the decomposition function expression is as follows:
[0039] ;
[0040] In the formula, represents the displacement time series along the sliding direction, represents the displacement periodic term, represents the intrinsic mode function obtained at the mth iteration, represents the total number of iterations, Represents the displacement trend term, which characterizes the residual component with the longest period.
[0041] As a preferred solution of the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellite described in the present invention, wherein: the displacement time series of the characteristic points of the landslide along the sliding direction is decomposed into displacement trend terms and displacement period terms to restore the long-term spatiotemporal deformation evolution of the landslide, and also includes: performing piecewise linear fitting on the displacement trend terms of ALOS-1 data and Sentinel-1 data, respectively calculating the displacement trend term rates of the two monitoring intervals by the least squares method, and using the mean of the displacement trend term rates of the two monitoring intervals to characterize and calculate the displacement along the sliding direction during the monitoring blank period, connecting the displacement period terms of the two monitoring intervals according to the displacement along the sliding direction during the monitoring blank period, and restoring the long-term spatiotemporal deformation evolution of the landslide.
[0042] Beneficial effects of the present invention:
[0043] The invention collects multi-platform and multi-band SAR satellite data covering the landslide, and performs landslide deformation inversion according to the wavelength and spatial resolution characteristics of the SAR satellite data. The landslide deformation inversion includes the line-of-sight time series deformation inversion of ALOS-1 data and Sentinel-1 data and the azimuth deformation rate inversion of TerraSAR-X data. The line-of-sight deformation rate and the line-of-sight displacement time series are generated according to the line-of-sight time series deformation inversion, and the azimuth deformation rate is generated according to the azimuth deformation rate inversion. The influence of vegetation decoherence is reduced through different inversion methods, and the inversion accuracy of azimuth deformation is improved. The azimuth deformation rate and the line-of-sight deformation rate in the overlapping monitoring interval are combined to establish an observation equation. , and reconstructed the average three-dimensional deformation rate of the landslide, which improved the monitoring dimension; the velocity vector equation was established through the reconstructed average three-dimensional deformation rate, the average velocity vector of each landslide pixel in the geographic coordinate system was calculated, and the sliding direction of each landslide pixel in the monitoring interval was obtained; the line-of-sight displacement time series of each landslide pixel was projected one by one to the sliding direction in the monitoring interval, and the displacement time series of the landslide along the sliding direction was obtained, which effectively solved the problem of inconsistent revisit periods of SAR satellite data on different platforms; the displacement time series of the landslide feature points along the sliding direction was decomposed into displacement trend terms and displacement period terms, and the long-term spatiotemporal deformation evolution of the landslide was restored, providing a more accurate and reliable technical reference for the deformation inversion of slow-moving landslides covered with vegetation.
[0044] By combining multi-platform SAR satellite data, the average three-dimensional deformation rate inversion of the landslide was achieved, and the optimal sliding direction of each landslide pixel was determined. This allowed us to unify the benchmark, project the line-of-sight displacement time series under different SAR satellite observation geometries to the same optimal sliding direction, and reconstruct the long-term spatiotemporal deformation evolution of the landslide. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:
[0046] Figure 1 It is a method flow chart of the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellites of the present invention;
[0047] Figure 2It is a spatiotemporal baseline diagram of multi-platform SAR satellite data of the reservoir area landslide long-term spatiotemporal deformation monitoring method based on multi-platform SAR satellites of the present invention, wherein (a) and (b) are interferometric baseline connection diagrams of ALOS-1 data and Sentinel-1 data processed by SBAS, and (c) represents the offset-to-baseline connection diagram of TerraSAR-X data processed by pixel offset tracking and stacking;
[0048] Figure 3 The three-dimensional deformation inversion result diagram of the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellite of the present invention, wherein (a) is the east-west velocity, (b) is the north-south velocity, (c) is the vertical velocity, and (d) is the horizontal velocity vector and the L2 norm of the three-dimensional velocity;
[0049] Figure 4 It is a spatiotemporal inversion diagram of landslide displacement of the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellites of the present invention;
[0050] Figure 5 This is a diagram of the long-term displacement inversion results of the landslide feature points of the reservoir area landslide long-term spatiotemporal deformation monitoring method based on multi-platform SAR satellites of the present invention. DETAILED DESCRIPTION
[0051] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.
[0052] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.
[0053] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.
[0054] Example 1
[0055] In this embodiment, a method flow chart of a method for long-term spatiotemporal deformation monitoring of landslides in a reservoir area based on multi-platform SAR satellites is provided, such as Figure 1 As shown in the figure, the long-term spatiotemporal deformation monitoring method of reservoir landslide based on multi-platform SAR satellite includes:
[0056] S1. Collect multi-platform and multi-band SAR satellite data covering the landslide, and perform landslide deformation inversion based on the wavelength and spatial resolution characteristics of the SAR satellite data. The landslide deformation inversion includes the line-of-sight time series deformation inversion of ALOS-1 data and Sentinel-1 data and the azimuth deformation rate inversion of TerraSAR-X data. The line-of-sight deformation rate and line-of-sight displacement time series are generated based on the line-of-sight time series deformation inversion, and the azimuth deformation rate is generated based on the azimuth deformation rate inversion.
[0057] Taking two reservoir landslides along the Yangtze River in the Three Gorges Reservoir area of Fengjie County, Chongqing, China as examples, the two reservoir landslides are Outang landslide and Xinpu landslide. 17 scenes of L-band ascending ALOS-1 data from 2007 to 2010, 172 scenes of C-band ascending Sentinel-1 data from 2015 to 2021, and 12 scenes of X-band ascending TerraSAR-X high-resolution focusing mode data from 2019 to 2021 were collected as examples. For the longer-wavelength ALOS-1 data and Sentinel-1 data, the SBAS algorithm is used to perform line-of-sight temporal deformation inversion; for the X-band high-spatial resolution TerraSAR-X data, the resolution here is 1 meter, and the InSAR method is prone to rapid decoherence. Therefore, the azimuth deformation rate inversion is performed through the pixel offset tracking algorithm, and multiple offset pairs are fused through the stacking algorithm to improve the accuracy of the azimuth deformation rate inversion.
[0058] Specific processing of multi-platform and multi-band SAR satellite data is as follows: Figure 1 As shown in Figure 1, the ALOS-1 data and Sentinel-1 data are screened according to the preset space-time baseline threshold to generate an interference pair combination, and the interference pair combination is inverted through SBAS to obtain the line-of-sight deformation rate and line-of-sight displacement time series. The interference pair baseline connection diagram of ALOS-1 data and Sentinel-1 data processed by SBAS is shown in Figure 1. Figure 2 As shown in (a) and (b) in Figure 1, the TerraSAR-X data are screened according to the preset spatiotemporal baseline threshold to generate offset pair combinations, and the offset pair combinations are processed by pixel offset tracking inversion and stacking to obtain the azimuthal deformation rate. A total of 35 offset pairs are used in the pixel offset tracking inversion and stacking process of TerraSAR-X data. The offset pair baseline connection diagram is shown in Figure 2 As shown in (c), Figure 2 The horizontal axis represents time, and the vertical axis represents the vertical baseline. Figure 2 Each black triangle in represents a SAR image, and the large gray triangles formed by the lines represent the common main image used for SAR image registration. The function expression of the stacking process is as follows:
[0059] ;
[0060] In the formula, represents the azimuth average deformation rate of TerraSAR-X data, Indicates The pixel offset tracking inversion results of the offset pairs along the azimuth direction, Indicates The time span of the offset pairs, Indicates the number of offset pairs.
[0061] Considering that the accumulation of displacement terms obtained by tracking and inverting the pixel offset of the offset pair along the azimuth direction in the stacking process changes linearly and is proportional to the number of offset pairs , while the background noise increases only with the square root of the number of offset pairs and is proportional to , so the function expression of the inversion error of the azimuthal average deformation rate is as follows:
[0062] ;
[0063] In the formula, represents the inversion error of the average deformation rate in the azimuth direction, Indicates The pixel offsets of each offset pair along the azimuth direction track the error in the inversion result.
[0064] What needs to be explained is: is about 0.1 m, when the time spans of all offset pairs are approximately equal, Equivalent to Therefore, after stacking, the pixel offset tracking inversion can theoretically Reduce to In this embodiment, the inversion error of the average deformation rate in the azimuth direction is approximately m / year.
[0065] The multi-platform and multi-band SAR satellite data include not only the data in the orbit raising direction in this embodiment, but also the data in different orbit directions, such as data in both the orbit raising direction and the orbit descending direction. However, it must be ensured that there are two or more valid independent observation results, such as at least one azimuth deformation and one line of sight deformation, or one orbit raising line of sight deformation and one orbit descending line of sight deformation, etc.
[0066] S2. Combine the azimuth deformation rate and the line-of-sight deformation rate within the overlapping monitoring interval, establish the observation equation, and reconstruct the average three-dimensional deformation rate of the landslide.
[0067] Under the premise of parallel surface motion, the deformation rate in the line of sight of Sentinel-1 data and the deformation rate in the azimuth of TerraSAR-X data in the overlapping monitoring interval are combined, and the functional expression of the observation equation is established as follows:
[0068] ;
[0069] In the formula, Indicates the line-of-sight deformation rate of Sentinel-1 data, represents the azimuthal deformation rate of TerraSAR-X data, represents the sine function, represents the cosine function, represents the incident angle of Sentinel-1 data, Represents the satellite heading angle of Sentinel-1 data, represents the satellite heading angle of TerraSAR-X data, represents the east-west velocity, represents the north-south velocity, represents the vertical velocity, represents the landslide elevation obtained from the digital elevation model, represents the first-order derivative of the landslide elevation along the north-south direction, Represents the first-order derivative of the landslide elevation along the east-west direction.
[0070] It needs to be explained that the digital elevation model adopts the global SRTM-1 digital elevation model.
[0071] The above-mentioned overlapping monitoring period refers to 2019 to 2021. Figure 3 As shown, the area within the solid box represents the stable reference area, where (a) represents the east-west velocity, (b) represents the north-south velocity, (c) represents the vertical velocity, and (d) represents the horizontal velocity vector and the L2 norm of the three-dimensional velocity.
[0072] When there are three or more valid independent observation results, that is, when three or more of the ascending orbit line of sight, ascending orbit azimuth, descending orbit line of sight and descending orbit azimuth are selected or exist at the same time, the established observation equation is as follows:
[0073] ;
[0074] In the formula, represents the line-of-sight deformation rate of the ascending orbit, represents the line-of-sight deformation rate of the descending orbit, represents the azimuthal deformation rate of the ascending orbit, represents the azimuthal deformation rate of the descending orbit, represents the incident angle of the ascending orbit, represents the incident angle of the descending orbit, Indicates the satellite heading angle of orbit raising, represents the satellite heading angle of descending orbit, Represents the inversion error of the observation equation.
[0075] It needs to be explained that when there are three independent observations, is 0, then the observation equations are positive definite, and the three-dimensional rate can be obtained by linear algebra. When there are more than three independent observation results, is not equal to 0, then the observation equations are overdetermined, and the weighted least squares method is used to minimize The three-dimensional velocity can be obtained by solving the set of equations.
[0076] S3. The velocity vector equation is established through the reconstructed average three-dimensional deformation rate, the average velocity vector of each landslide pixel in the geographic coordinate system is calculated, and the optimal sliding direction of each landslide pixel in the monitoring area is obtained.
[0077] The function expression of the velocity vector equation is as follows:
[0078] ;
[0079] In the formula, represents the unit velocity vector of the landslide pixel, Represents the velocity vector of the landslide pixel in the geographic coordinate system, Represents the L2 norm of the three-dimensional rate.
[0080] Landslide pixel refers to each pixel resolution unit on the landslides in the two reservoir areas under monitoring, that is, each monitoring point on the landslides.
[0081] S4. Project the sight-direction displacement time series of each landslide pixel to the optimal sliding direction within the monitoring interval one by one to obtain the displacement time series of the landslide along the optimal sliding direction.
[0082] like Figure 4 As shown, it is the incremental displacement of the landslide along the optimal sliding direction every 60 days from 2017 to 2020. The functional expression of the displacement time series is as follows:
[0083] ;
[0084] In the formula, represents the displacement time series of the landslide along the optimal sliding direction, represents the time series of line-of-sight displacement of ALOS-1 data and Sentinel-1 data calculated by the SBAS algorithm, Unit vector representing the direction of the radar's line of sight.
[0085] Among them, the functional expression of the line of sight displacement time series is as follows:
[0086] ;
[0087] in, Indicates The time series of line-of-sight displacements of the interference pairs is obtained by inverting the line-of-sight time series deformation.
[0088] S5. Select the landslide pixels with significant deformation characteristics as landslide feature points, such as Figure 3 As shown in X1-X4 in , the displacement time series of these landslide characteristic points along the optimal sliding direction are decomposed into displacement trend terms and displacement period terms, which are then used to restore the long-term spatiotemporal deformation evolution of the landslide.
[0089] Since the deformation of landslides in the reservoir area is mainly affected by internal and external forces, the deformation of landslides under the influence of internal forces presents a displacement trend term with small fluctuations and long-term consistency, while the deformation of landslides under the influence of external forces presents a displacement period term with regular fluctuations, where external forces include rainfall and reservoir water level scheduling. In order to separate the long-term spatiotemporal deformation trend of landslides, the displacement time series of ALOS-1 data and Sentinel-1 data along the optimal sliding direction are decomposed by the empirical mode decomposition algorithm. The decomposed function expression is as follows:
[0090] ;
[0091] In the formula, represents the displacement time series along the optimal sliding direction, represents the displacement periodic term, represents the intrinsic mode function obtained at the mth iteration, represents the total number of iterations, Represents the displacement trend term, which characterizes the residual component with the longest period.
[0092] It needs to be explained that each Both represent a specific frequency and amplitude modulation component in the displacement time series along the optimal sliding direction, which increases with the number of iterations during the operation of the empirical mode decomposition algorithm. increase, The frequency will decrease accordingly.
[0093] The displacement trend terms of ALOS-1 data and Sentinel-1 data were piecewise linearly fitted, and the displacement trend term rates of the two monitoring intervals were calculated using the least squares method. The mean of the displacement trend term rates of the two monitoring intervals was used to represent the displacement rate along the optimal sliding direction during the monitoring blank period. The mean of the displacement trend term rates of the two monitoring intervals was multiplied by the length of the monitoring blank period to obtain the displacement along the optimal sliding direction during the monitoring blank period. The displacement periodic terms of the two monitoring intervals were connected based on this, and the long-term spatiotemporal deformation evolution of the landslide was restored.
[0094] On the basis of the above decomposition, piecewise linear fitting is performed on the displacement trend items of ALOS-1 data and Sentinel-1 data. The displacement trend item rates of the monitoring interval from 2007 to 2010 in ALOS-1 data and the monitoring interval from 2015 to 2021 in Sentinel-1 data are solved by the least squares method, and the mean of the displacement trend item rates of the two monitoring intervals is used to represent the displacement rate along the optimal sliding direction during the monitoring blank period from 2011 to 2015. The displacement rate is multiplied by the duration of the monitoring blank period to obtain the displacement along the optimal sliding direction during the monitoring blank period from 2011 to 2015. At the same time, the monitoring interval in the ALOS-1 data is from 2007 to 2010 to generate the displacement period term of the corresponding monitoring interval, and the monitoring interval in the Sentinel-1 data is from 2015 to 2021 to generate the displacement period term of the corresponding monitoring interval. The displacement along the optimal sliding direction during the monitoring blank period from 2011 to 2015 is used to connect the displacement period terms of the above two monitoring intervals to restore the long-term spatiotemporal deformation evolution of the landslide for up to 15 years, which is specifically expressed as follows: Figure 5 As shown in the figure, X1-X4 represent the above selected landslide characteristic points, the horizontal axis represents time, which refers to the sum of the time of the above monitoring interval and monitoring interval, up to 15 years, and the vertical axis represents the cumulative displacement along the best sliding direction. Figure 5 The monitoring interval in the monitoring is the monitoring blank period.
[0095] Working principle and its effect:
[0096] (1) The landslide deformation was inverted using multi-platform SAR satellite data, and appropriate inversion algorithms were selected according to the wavelength and spatial resolution characteristics of SAR satellite data. The influence of vegetation decoherence was minimized by performing InSAR processing on long-wavelength, high-penetration SAR data. The azimuthal deformation inversion was achieved by performing pixel offset tracking processing on high-resolution X-band SAR data, and the stacking algorithm was used to effectively improve the inversion accuracy of the azimuthal deformation, with an accuracy of up to centimeters.
[0097] (2) By combining the landslide deformation monitoring results with multi-platform SAR satellite data, the average three-dimensional displacement rate inversion of the landslide was achieved, which improved the monitoring dimension, revealed the true movement state of the landslide, and solved the problem that traditional InSAR algorithms cannot monitor north-south moving landslides.
[0098] (3) The optimal sliding direction of each landslide pixel is determined by averaging the three-dimensional displacement rate. The line-of-sight displacement time series under different satellite observation geometries can then be projected pixel by pixel to a unified reference direction, thus avoiding the cumbersome process of performing three-dimensional deformation inversion for each phase. This effectively solves the problem of inconsistent revisit periods of SAR satellite data from different platforms and the difficulty in obtaining multiple independent observations at the same time for three-dimensional time series deformation inversion.
[0099] (4) The optimal sliding direction displacement time series of different monitoring intervals were decomposed using empirical mode decomposition, and the displacement trend term was fitted with piecewise linear fitting, so as to connect the monitoring blank periods between SAR data of different platforms and restore the long-term spatiotemporal deformation evolution of the landslide. This method provides a more accurate and reliable technical reference for deformation inversion of slow-moving landslides covered by vegetation and for the prevention, control and assessment of landslide disasters in reservoir areas.
[0100] Example 2
[0101] In this embodiment, a computer device is provided, including a memory and a processor, the memory is used to store instructions, and the processor is used to execute the instructions, so that the computer device executes the steps of implementing the above-mentioned method for long-term spatiotemporal deformation monitoring of reservoir landslides based on multi-platform SAR satellites.
[0102] Example 3
[0103] In this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed, the steps of the above-mentioned method for long-term spatiotemporal deformation monitoring of reservoir landslides based on multi-platform SAR satellites are implemented.
[0104] The computer-readable storage medium includes: a U disk, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and other media for storing program codes.
[0105] It should be noted that the above embodiments are only used to illustrate the present invention rather than to limit it; although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that without departing from the spirit and scope of the present invention, deduction, deformation and modification of the specific implementation methods of the present invention will fall within the scope defined by the claims of the present invention and also belong to the protection scope of the present invention.
Claims
1. A long-term spatiotemporal deformation monitoring method for reservoir landslides based on multi-platform SAR satellites, characterized in that: include: Collect multi-platform and multi-band SAR satellite data covering the landslide, and perform landslide deformation inversion according to the wavelength and spatial resolution characteristics of the SAR satellite data. The landslide deformation inversion includes the line-of-sight time series deformation inversion of ALOS-1 data and Sentinel-1 data and the azimuth deformation rate inversion of TerraSAR-X data. The line-of-sight deformation rate and line-of-sight displacement time series are generated according to the line-of-sight time series deformation inversion, and the azimuth deformation rate is generated according to the azimuth deformation rate inversion. The azimuth deformation rate and the sightline deformation rate in the overlapping monitoring interval are combined to establish the observation equation and reconstruct the average three-dimensional deformation rate of the landslide. A velocity vector equation is established by reconstructing the average three-dimensional deformation rate, the average velocity vector of each landslide pixel in the geographic coordinate system is calculated, and the sliding direction of each landslide pixel in the monitoring interval is obtained; Projecting the sight-direction displacement time series of each landslide pixel to the sliding direction within the monitoring interval one by one, and obtaining the displacement time series of the landslide along the sliding direction; The displacement time series of the characteristic points of the landslide along the sliding direction is decomposed into displacement trend terms and displacement period terms to restore the long-term spatiotemporal deformation evolution of the landslide.
2. The method for long-term spatiotemporal deformation monitoring of reservoir landslides based on multi-platform SAR satellites as claimed in claim 1, characterized in that: The SAR satellite data include data covering different orbital directions of the same landslide and data in a single orbital direction, and two or more independent observation results are maintained.
3. The method for long-term spatiotemporal deformation monitoring of reservoir landslides based on multi-platform SAR satellites as claimed in claim 2 is characterized in that: The azimuth deformation rate inversion of the TerraSAR-X data includes: performing azimuth deformation rate inversion on the TerraSAR-X data by combining pixel offset tracking and stacking processing to generate azimuth deformation rate, wherein the function expression of the stacking processing is as follows: ; In the formula, represents the azimuth average deformation rate of TerraSAR-X data, Indicates The pixel offset tracking inversion results of the offset pairs along the azimuth direction, Indicates The time span of the offset pairs, Indicates the number of offset pairs; In the stacking process, the accumulation of displacement terms obtained by tracking and inverting the pixel offset of the offset pair along the azimuth direction changes linearly and is proportional to the number of offset pairs. , while the inverted background noise increases only with the square root of the number of offset pairs and is proportional to , the functional expression of the inversion error of the azimuthal average deformation rate is as follows: ; In the formula, represents the inversion error of the average deformation rate in the azimuth direction, Indicates The pixel offsets of each offset pair along the azimuth direction track the error in the inversion result.
4. The method for long-term spatiotemporal deformation monitoring of reservoir landslides based on multi-platform SAR satellites as claimed in claim 3 is characterized in that: The establishment rules of the observation equation include: If there are three or more independent observations, an observation equation is established; If there are two independent observation results, the observation equation is established in combination with the parallel motion of the surface.
5. The method for long-term spatiotemporal deformation monitoring of reservoir landslides based on multi-platform SAR satellites as claimed in claim 4 is characterized in that: There are three or more independent observations, and the established observation equation is as follows: ; In the formula, represents the line-of-sight deformation rate of the ascending orbit, represents the line-of-sight deformation rate of the descending orbit, represents the azimuthal deformation rate of the ascending orbit, represents the azimuthal deformation rate of the descending orbit, represents the incident angle of the ascending orbit, represents the incident angle of the descending orbit, Indicates the satellite heading angle of orbit raising, represents the satellite heading angle of descending orbit, represents the inversion error of the observation equation; The functional expression of the observation equation established in combination with the parallel motion of the earth's surface is as follows: ; In the formula, Indicates the line-of-sight deformation rate of Sentinel-1 data, represents the azimuthal deformation rate of TerraSAR-X data, represents the sine function, represents the cosine function, represents the incident angle of Sentinel-1 data, Represents the satellite heading angle of Sentinel-1 data, represents the satellite heading angle of TerraSAR-X data, represents the east-west velocity, represents the north-south velocity, represents the vertical velocity, represents the landslide elevation obtained from the digital elevation model, represents the first-order derivative of the landslide elevation along the north-south direction, Represents the first-order derivative of the landslide elevation along the east-west direction.
6. The method for long-term spatiotemporal deformation monitoring of reservoir landslides based on multi-platform SAR satellites as claimed in claim 5, characterized in that: The function expression of the velocity vector equation is as follows: ; In the formula, represents the unit velocity vector of the landslide pixel, Represents the velocity vector of the landslide pixel in the geographic coordinate system, Represents the L2 norm of the three-dimensional rate.
7. The method for long-term spatiotemporal deformation monitoring of reservoir landslides based on multi-platform SAR satellites as claimed in claim 6, characterized in that: The functional expression of the displacement time series of the landslide along the sliding direction is as follows: ; In the formula, represents the displacement time series of the landslide along the sliding direction, represents the time series of line-of-sight displacement of ALOS-1 data and Sentinel-1 data calculated by the SBAS algorithm, The unit vector representing the radar line of sight direction; Among them, the expression of the time series of line of sight displacement is: ,in, Indicates The time series of line-of-sight displacements of the interference pairs is obtained by inverting the line-of-sight time series deformation.
8. The method for long-term spatiotemporal deformation monitoring of reservoir landslides based on multi-platform SAR satellites as claimed in claim 7, characterized in that: The method of decomposing the displacement time series of the characteristic points of the landslide along the sliding direction into a displacement trend term and a displacement period term to restore the long-term spatiotemporal deformation evolution of the landslide includes: decomposing the displacement time series of the ALOS-1 data and the Sentinel-1 data along the sliding direction respectively by using an empirical mode decomposition algorithm, and the decomposition function expression is as follows: ; In the formula, represents the displacement periodic term, represents the intrinsic mode function obtained at the mth iteration, represents the total number of iterations, Represents the displacement trend term, which characterizes the residual component with the longest period.
9. The method for long-term spatiotemporal deformation monitoring of reservoir landslides based on multi-platform SAR satellites as claimed in claim 8, characterized in that: The method of decomposing the displacement time series of the characteristic points of the landslide along the sliding direction into displacement trend items and displacement period items to restore the long-term spatiotemporal deformation evolution of the landslide also includes: performing piecewise linear fitting on the displacement trend items of the ALOS-1 data and the Sentinel-1 data, respectively calculating the displacement trend item rates of the two monitoring intervals by the least squares method, and using the mean of the displacement trend item rates of the two monitoring intervals to characterize and calculate the displacement along the sliding direction during the monitoring blank period, connecting the displacement period items of the two monitoring intervals according to the displacement along the sliding direction during the monitoring blank period, and restoring the long-term spatiotemporal deformation evolution of the landslide.
Citation Information
Patent Citations
A method for monitoring and early warning of landslide deformation based on SAR data
CN111474544B
Landslide deformation monitoring method
CN117872350A
Landslide deformation monitoring and early warning method based on SAR data
CN111474544A
Coastal dam safety monitoring and early warning method and system
CN117516636A
Method for extracting motion errors of a carrier carrying a coherent imaging radar system from raw radar data and device for carrying out the method
DE4311754C1
