A method and system for identifying landslide geological hazards in reservoir areas

Abstract

The present application relates to the field of landslide geological identification technology, and specifically to a method and system for identifying landslide geological hazards in reservoir areas. The method comprises: collecting SAR images and DEM data of the reservoir area at various time points; combining SAR images of adjacent time points into a master-slave SAR image, obtaining an interferogram of the master-slave SAR image, and dividing the interferogram of the master-slave SAR image into a decoherent water area, a terrain transition area, and a high-coherence stable area; determining the terrain phase gradient of each pixel point in the interferogram to obtain the terrain gradient amplitude of each pixel point, correcting the objective function of a least squares phase unwrapping algorithm, and obtaining the interferometric phase after phase unwrapping; determining the elevation information corresponding to the interferogram, calculating the average deformation rate of each pixel point, and identifying the potential landslide risk in the reservoir area corresponding to each pixel point. This improves the accuracy of landslide risk identification in reservoir areas.

Description

Technical Field

[0001] The present application relates to the technical field of landslide geological identification, and specifically to a method and system for identifying landslide geological disasters in reservoir areas. Background Art

[0002] Synthetic Aperture Radar (Synthetic Aperture Radar) interferometry (D-InSAR) extracts three-dimensional surface information by analyzing phase differences between two SAR images of the same area. This allows for high-precision monitoring of surface deformation and is widely used in fields such as geological disaster early warning and geographic deformation analysis. D-InSAR technology provides critical data support for early landslide warning in reservoir areas. Phase unwrapping, a core step in D-InSAR, directly impacts the quality of the interferogram and the accuracy of deformation inversion.

[0003] However, in the identification of landslide hazards in reservoir areas, reservoir areas are usually located in canyon and mountainous areas with complex terrain, and are affected by periodic water storage and drainage, resulting in complex surface deformation patterns in the coastal areas. During the water storage period, the SAR signals in the water surface area and the coastal area are severely decoherent. During the drainage period, the steep slope terrain of the reservoir bank introduces high-frequency phase gradients. The traditional least squares unwrapping method will misjudge the terrain phase as a deformation signal when processing areas with large phase gradient changes. The canyon terrain with complex gradient changes will also affect the accuracy of the deformation phase obtained by the least squares unwrapping algorithm, which in turn leads to low accuracy and reliability of differential interferometry processing, reducing the effect of landslide risk identification in reservoir areas. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide a method and system for identifying landslide geological hazards in reservoir areas. The technical solutions adopted are as follows:

[0005] In a first aspect, an embodiment of the present application provides a method for identifying landslide geological hazards in a reservoir area, the method comprising the following steps:

[0006] Collect SAR images and DEM data of the reservoir area at each time point;

[0007] SAR images at adjacent time points are combined into a master and slave SAR image, and the interferogram of the master and slave SAR images is obtained. Based on the magnitude relationship between the coherence coefficients of the pixels of the master and slave SAR images, the interferogram of the master and slave SAR images is divided into a decoherent water area, a terrain transition area, and a high coherence stable area.

[0008] The terrain phase of the reservoir area is determined by the geometric parameters of the DEM data and SAR image acquisition. The terrain phase gradient of each pixel point in the interference pattern is determined using the terrain phase, and the terrain gradient amplitude of each pixel point is obtained.

[0009] The gradient amplitude of the unwrapped phase of each pixel in the interferogram is taken as an unknown term. According to the difference between the gradient amplitude of the entangled phase and the gradient amplitude of the unwrapped phase of each pixel in the interferogram, and the coherence coefficient deviation of the pixel points in the decoherent water area, the terrain transition area and the high coherence stable area, the difference is weighted. Combined with the terrain gradient amplitude, the objective function of the least squares phase unwrapping algorithm is modified to solve the unknown term and obtain the interference phase after phase unwrapping;

[0010] Based on the interference phase, the deformation phase of the interference pattern is obtained, and combined with the DEM data, the elevation information corresponding to the interference pattern is determined; according to the degree of change of the elevation information corresponding to the interference pattern at all time points, the average deformation rate of each pixel point is calculated, and the potential landslide risk of the reservoir area corresponding to each pixel point is identified.

[0011] In one embodiment, dividing the interferograms of the master and slave SAR images into a decoherent water area, a terrain transition area, and a high coherence stable area includes:

[0012] Using a multi-threshold segmentation algorithm, the coherence coefficients of all pixels in the interference pattern are segmented to obtain a first threshold and a second threshold, wherein the first threshold is less than the second threshold;

[0013] The area where the pixels with a coherence coefficient less than the first threshold in the interference map are located is regarded as the incoherent water area, the area where the pixels with a coherence coefficient greater than or equal to the first threshold and less than or equal to the second threshold in the interference map are located is regarded as the terrain transition area, and the area where the pixels with a coherence coefficient greater than the second threshold in the interference map are located is regarded as the high coherence stable area.

[0014] In one embodiment, the topographic phase gradient includes a topographic phase gradient in a horizontal direction and a topographic phase gradient in a vertical direction of each pixel point.

[0015] In one embodiment, the terrain gradient amplitude is calculated as follows:

[0016] Where, is the pixel point in the interference pattern The terrain gradient amplitude, is the pixel point in the interference pattern The terrain phase gradient in the horizontal direction, is the pixel point in the interference pattern The terrain phase gradient in the vertical direction, is the horizontal and vertical coordinates of the pixel point.

[0017] In one embodiment, weighting the difference comprises:

[0018] When any pixel point in the interference map is located in the decoherent water area, the weight of the difference corresponding to the pixel point is 0; when any pixel point in the interference map is located in the terrain transition area, the mean of the first threshold and the second threshold is calculated, and the weight of the difference corresponding to the pixel point is the sum of the mean and the second threshold. The product of , where e is a natural constant; when any pixel point in the interference pattern is located in a high coherence stable region, the weight of any pixel point corresponding to the difference is the coherence coefficient of any pixel point.

[0019] In one embodiment, the objective function is expressed as:

[0020] Where, is the objective function of the interference pattern, is the pixel point in the interference pattern The weight corresponding to the area, is the size of the interference pattern, is the pixel point in the interference pattern The gradient amplitude of the unwrapped phase, is the pixel point in the interference pattern The gradient amplitude of the winding phase, is the preset smoothing factor, is the pixel point in the interference pattern The magnitude of the terrain gradient.

[0021] In one embodiment, obtaining the deformation phase of the interference pattern includes:

[0022] The terrain phase, the atmospheric phase and the noise phase in the interference phase after phase unwrapping are determined, and the terrain phase, the atmospheric phase and the noise phase are removed from the interference phase to obtain the deformation phase.

[0023] In one embodiment, identifying the potential landslide risk of the reservoir area corresponding to each pixel point includes:

[0024] The average deformation rates of all pixels are arranged in descending order, and the average of the first preset percentage of average deformation rates is used as the judgment threshold. The relationship between the average deformation rate of all pixels and the judgment threshold is used to identify the potential landslide risk in the reservoir area corresponding to each pixel.

[0025] In one embodiment, the reservoir area corresponding to the pixel points whose average deformation rate is greater than the judgment threshold is regarded as the landslide risk area.

[0026] In the second aspect, an embodiment of the present application also provides a reservoir area landslide geological disaster identification system, including a memory, a processor, and a computer program stored in the memory and running on the processor, and when the processor executes the computer program, it implements the steps of any one of the above methods.

[0027] This application has at least the following beneficial effects:

[0028] This application mainly addresses the limitations of the traditional least squares phase unwrapping algorithm, proposes an improved least squares phase unwrapping algorithm based on regional gradient adaptive regularization, and constructs an objective function for phase unwrapping. The data fitting term weight of the objective function is dynamically adjusted by the coherence and terrain gradient of different regions in the interferogram. The gradient phase of the high coherence region is dominant, the phase noise caused by the incoherence region is suppressed, and the accuracy of phase unwrapping in the interferogram is improved. In order to avoid the phase distortion problem caused by complex terrain gradient changes in the canyon and high gradient changes along the reservoir coast, a regularization term is designed based on the terrain gradient characteristics. The terrain gradient is used to guide phase unwrapping, separate the terrain phase and the deformation phase, avoid the terrain gradient with strong changes from being misjudged as deformation signals, improve the recognition ability of detailed deformation phase, and avoid the problem of deformation phase distortion caused by the traditional least squares phase unwrapping algorithm due to decoherent signals in the water area, complex terrain information and gradient mutation characteristics in the coastal area. This further improves the accuracy of phase unwrapping in complex reservoir area scenarios, improves the quality of SAR image differential interferograms, provides a reliable data basis for the detection of landslide disasters, and thus improves the accuracy of landslide disaster identification in reservoir areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0030] Figure 1 A flowchart of a method for identifying landslide geological hazards in a reservoir area provided in one embodiment of the present application;

[0031] Figure 2 Construct a flowchart for the objective function. DETAILED DESCRIPTION

[0032] To further illustrate the technical means and effectiveness of this application to achieve the intended invention objectives, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail a method and system for identifying landslide geological hazards in reservoir areas proposed in this application, including its specific implementation, structure, features, and effectiveness. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0033] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0034] The following describes in detail a method and system for identifying landslide geological hazards in reservoir areas provided by this application with reference to the accompanying drawings.

[0035] See also Figure 1 , which shows a flowchart of a method for identifying landslide geological hazards in a reservoir area provided by an embodiment of the present application, the method comprising the following steps:

[0036] Landslide hazard identification in reservoir areas is a common problem. Reservoirs are often located in canyon-mountainous areas with complex terrain, and impoundment causes changes in the seepage pressure of rock and soil along the reservoir banks, making landslides more likely to occur. Furthermore, reservoirs periodically store and discharge water. During periods of high water levels, the water surface exhibits strong scattering characteristics, leading to severe decoherence of SAR signals between the water area and the coastal area. Low water levels during drainage expose steep slopes along the reservoir banks, creating abrupt phase gradients that interfere with deformation signals in the SAR signals.

[0037] Due to its high reflectivity and independence from terrain information, the water surface area exhibits decoherence with the deformation phase. The traditional least squares phase unwrapping algorithm uses the same weight to calculate the unwrapped phase for the gradient residual calculation of global pixel points, resulting in the phase noise of the decoherent water surface area being propagated to the terrain area with high coherence in the global residual gradient calculation in the least squares phase unwrapping algorithm, causing errors in the deformation phase estimation of the terrain.

[0038] In addition, the reservoir water level drops sharply during the drainage period, and the area along the reservoir shows large terrain gradient changes. The traditional least squares phase unwrapping algorithm will misjudge the terrain phase as a deformation signal when processing areas with large phase gradient changes. The canyon terrain with complex gradient changes will also affect the accuracy of the least squares phase unwrapping algorithm.

[0039] To address the above issues, this application performs phase unwrapping on the differential interferometry processing of SAR in reservoirs through an improved least squares phase unwrapping algorithm based on regional gradient adaptive regularization. The weight values ​​of the pixel gradient residuals in different areas of the reservoir area are adaptively adjusted in the global phase gradient estimation function based on the coherence analysis and gradient changes in different areas of the reservoir area. In addition, a terrain gradient regularization term is added in combination with the external terrain gradient to improve the recognition ability of detailed deformation phases and avoid the problem of deformation phase distortion caused by the traditional least squares phase unwrapping algorithm due to incoherent signals in the water area, complex terrain information and gradient mutation characteristics in the coastal area. Specifically:

[0040] S1, collect SAR images and DEM data of the reservoir area at each time point, and combine the SAR images of adjacent time points into a master-slave SAR image.

[0041] This embodiment uses satellites to acquire SAR image data of a reservoir area at various time points. Baseline estimation is performed based on SAR image data acquired at different time points for the same reservoir area, and short baseline data is obtained to form a multi-temporal SAR image dataset. Specifically, multi-temporal SAR image data is SAR image data of the same reservoir area acquired at different time points. This embodiment sets the acquisition interval to 15 days, and the total acquisition duration to one year. This can be set by the implementer based on actual circumstances, and this embodiment does not impose any restrictions. SAR image data acquired at adjacent time points are combined to form a pair of master and slave SAR images for subsequent interferogram generation. Specifically, the SAR image at the earlier time point in the temporal sequence of adjacent time points is used as the master SAR image, and the SAR image at the later time point in the temporal sequence is used as the slave SAR image.

[0042] Radiometric correction is performed on short-baseline SAR datasets to improve data consistency between SAR images and real-world objects. Amplitude cross-correlation is used to register multi-temporal SAR images, ensuring pixel-level alignment between the master and slave SAR images. Radiometric correction and amplitude cross-correlation registration are existing technologies, and the detailed process is not detailed here.

[0043] Secondly, high-precision DEM data of the reservoir area was acquired via satellite. The DEM data was resampled using bilinear interpolation to match the resolution to the SAR imagery, ensuring alignment between the DEM and SAR images. Terrain correction was then performed to generate simulated terrain phases to facilitate their removal in the subsequent differential interferometry (DII) process. Bilinear interpolation is an established technique, and its detailed process is omitted here.

[0044] S2, obtaining the interferogram of the master and slave SAR images, and dividing the interferogram of the master and slave SAR images into a decoherent water area, a terrain transition area, and a high coherence stable area based on the magnitude relationship between the coherence coefficients of the pixels of the master and slave SAR images.

[0045] First, based on any pair of master and slave SAR images, an interferogram of the master and slave SAR images is obtained. The acquisition of the interferogram is a well-known technique, and the specific process is not described in detail.

[0046] This embodiment uses an arbitrary interferogram as an example for analysis. The coherence coefficients of pixels at the same location in the master and slave SAR images corresponding to the arbitrary interferogram are calculated. A multi-threshold segmentation algorithm is then used to segment the coherence coefficients of all pixels in the arbitrary interferogram, obtaining a first threshold and a second threshold, where the first threshold is less than the second threshold. In this embodiment, the multi-threshold segmentation algorithm uses the OTSU algorithm. Implementers may choose other feasible existing multi-threshold segmentation algorithms, and this embodiment does not limit this. The calculation of the coherence coefficient is well-known, and the specific process is not detailed here.

[0047] Secondly, the area where the pixels in the interference map with a coherence coefficient less than the first threshold are located is regarded as the incoherent water area, the area where the pixels in the interference map with a coherence coefficient greater than or equal to the first threshold and less than or equal to the second threshold are located is regarded as the terrain transition area, and the area where the pixels in the interference map with a coherence coefficient greater than the second threshold are located is regarded as the high coherence stable area.

[0048] By dividing the interference pattern into three regions: high-coherence stable regions, terrain transition regions, and incoherent water regions, a regional gradient adaptive regularization constraint term is constructed for each region based on the DEM terrain gradient characteristics. A second-order smoothing constraint is used in the high-coherence stable region to maintain deformation phase continuity. A terrain gradient matching term is introduced in the reservoir bank area belonging to the terrain transition region to separate the terrain and deformation phase. The incoherent water region is excluded from the optimization of the global residual gradient function to avoid interference caused by incoherent signals. Finally, a least squares objective function is constructed based on the data fitting term and the terrain gradient regularization term. The optimal unwrapped phase is solved by minimizing the objective function, and a high-precision reservoir deformation phase is output. This avoids the deformation phase distortion problem caused by incoherent water signal interference, steep slope terrain interference, and sudden changes in reservoir bank gradients in traditional algorithms.

[0049] Coherence reflects changes in features in both master and slave SAR images. High coherence typically corresponds to stable, resistant areas, such as mountains, rock, and soil in reservoir areas. Transition zones, defined as those between high and low coherence, exhibit minimal changes in both master and slave SAR images. For example, moderate coherence is typically observed in areas of slow surface deformation during the initial stages of a landslide disaster, as well as along the reservoir coast during the impoundment and drainage periods. Low coherence typically corresponds to areas experiencing drastic changes in both master and slave SAR images. Reservoir waters exhibit strong low coherence during the impoundment period, and high low coherence occurs during the impoundment and drainage periods due to water level fluctuations. Furthermore, the strong reflectivity of the water surface results in strong scattering in SAR images of features acquired from radar signals reflected from the water surface. Even when both master and slave SAR images are acquired during the impoundment period, high low coherence is observed in the water surface. Low high coherence in the water surface can lead to random phase noise during phase unwrapping of the interferogram, resulting in errors in the global phase gradient estimation.

[0050] S3, determine the terrain phase of the reservoir area through the geometric parameters of DEM data and SAR image acquisition; use the terrain phase to determine the terrain phase gradient of each pixel point in the interference pattern, and obtain the terrain gradient amplitude of each pixel point.

[0051] This embodiment calculates the terrain phase of the reservoir area by combining the DEM data after pixel-level registration with the geometric parameters of SAR image acquisition, including vertical baseline, incident angle, wavelength, and slant range. , terrain phase The calculation method of is an existing technology, and its specific process will not be repeated here.

[0052] Get terrain phase Afterwards, the central difference method is used to calculate the pixel point corresponding to any interference pattern. The topographic phase gradient in the horizontal and vertical directions and , and then get the pixel point The terrain gradient amplitude , where is the pixel point in the interference pattern The terrain gradient amplitude, is the pixel point in the interference pattern The terrain phase gradient in the horizontal direction, is the pixel point in the interference pattern The terrain phase gradient in the vertical direction, are the horizontal and vertical coordinates of the pixel point. The central difference method is an existing technology and its specific process will not be repeated here.

[0053] S4, taking the gradient amplitude of the unwrapped phase of each pixel point in the interference pattern as the unknown term, according to the difference between the gradient amplitude of the entangled phase and the gradient amplitude of the unwrapped phase of each pixel point in the interference pattern, and using the coherence coefficient deviation of the pixel points in the decoherent water area, the terrain transition area and the high coherence stable area, the difference is weighted, combined with the terrain gradient amplitude, the objective function of the least squares phase unwrapping algorithm is corrected, the unknown term is solved, and the interference phase after phase unwrapping is obtained.

[0054] The topographic gradient characteristics within the reservoir area exhibit significant variability, primarily influenced by canyon topography and periodic water storage and drainage activities. Canyon topography possesses complex topographic information, with high topographic phase gradient amplitudes, which appear as dense streaks or jumps in the interferogram and are easily confused with deformation signals. Furthermore, the reservoir banks have steep topographic slopes. Rising water levels during the impoundment period can mask some of the steep slopes, while falling water levels during the drainage period lead to sudden changes in topographic gradient amplitudes. This prolonged immersion of rock and soil in water and changes in its osmotic pressure make the reservoir banks susceptible to landslides. Landslides are characterized by low-frequency deformation gradients, and the high-frequency topographic gradient amplitude changes brought about by the drainage period can mask the true landslide deformation signals.

[0055] Based on the above analysis, this embodiment constructs the objective function of phase unwrapping based on the least squares phase unwrapping algorithm improved by regional gradient adaptive regularization, and solves the optimal unwrapping phase by minimizing the objective function.

[0056] The goal of the least squares phase unwrapping algorithm is to construct an objective function to minimize the difference between the wrapped phase gradient and the unwrapped phase gradient. The entangled phase between the two is restored to a continuous absolute phase. This is essentially a global optimization problem. However, in the phase unwrapping process of SAR differential interferometry in reservoir areas, the interferogram in the reservoir area has limitations due to the complex terrain structure of the canyon area and the periodic water storage and drainage activities of the reservoir.

[0057] The objective function of the least-squares phase unwrapping algorithm optimizes the phase gradients of all pixels in the interferogram to minimize the overall error. This objective function often applies the same weight to all pixels to solve for the phase gradient. However, the incoherent nature of the reservoir's large water area causes random gradients in the incoherent region to interfere with the unwrapped phase, resulting in a significant discrepancy between the obtained unwrapped phase and the true phase. Furthermore, strong gradients along the reservoir's shoreline can mask signals from deformation areas with lower gradients, leading to situations where these strong gradients are misinterpreted as deformation signals, resulting in errors in the unwrapped phase.

[0058] To address the above issues, the weight of the data fitting term of the objective function is dynamically adjusted in combination with the coherence of different regions in the interference pattern and the terrain gradient. The gradient phase of the high coherence region is dominant, the phase noise caused by the decoherence region is suppressed, and the accuracy of the phase unwrapping in the interference pattern is improved. In order to avoid the phase distortion problem caused by the complex terrain gradient changes in the canyon and the high gradient changes along the reservoir coast, a regularization term is designed based on the terrain gradient characteristics, and the terrain gradient is used to guide the phase unwrapping, separating the terrain phase from the deformation phase. The objective function of the least squares phase unwrapping algorithm improved based on regional gradient adaptive regularization in this embodiment is:

[0059] Where, is the objective function of the interference pattern, is the pixel point in the interference pattern The weight corresponding to the area, is the size of the interference pattern, indicating the coexistence of pixels, is the pixel point in the interference pattern The gradient amplitude of the unwrapped phase, is the item to be solved, is the pixel point in the interference pattern The gradient amplitude of the winding phase is obtained by differentially calculating the gradient phase of the pixel point in the interference pattern in the horizontal and vertical directions. is the preset smoothing factor, is the pixel point in the interference pattern The magnitude of the terrain gradient. is a gradient regularization term, the purpose of which is to assist the separation of terrain gradient phase and deformation gradient phase when unwrapping the phase of complex gradient change areas through terrain gradient to avoid phase distortion. In this embodiment, , the implementer can set it according to the actual situation, and this embodiment does not limit it. The objective function construction flow chart is as follows Figure 2 shown.

[0060] It should be noted that It represents the dynamic weight of the data fitting term, which can dynamically adjust the weight of the data fitting term according to the coherence of the pixels in different areas of the interference pattern. Specifically: when the pixel in the interference pattern Located in the decoherent water area, the pixel Corresponding The weight value is 0; when the pixel point in the interference pattern Located in the terrain transition area, calculate the average of the first threshold and the second threshold, the pixel point Corresponding The weight value is the mean and The product of, where e is a natural constant; when the pixel point in the interference pattern Located in the high coherence stable area, the pixel Corresponding The weight value is the pixel point The coherence coefficient.

[0061] It should be understood that when a pixel point belongs to a high-coherence terrain area, the coherence coefficient of the pixel point is used as the weight. The higher the coherence coefficient, the stronger the phase unwrapping effect of the pixel point is, and the true deformation signal is retained. When the pixel point belongs to the terrain transition zone, which contains the slight surface deformation in the early stage of landslide disasters and the reservoir coastal area during the drainage period, in order to further balance the weights of terrain and deformation signals, the As the terrain gradient decays, the interference of terrain phase is reduced. When phase unwrapping is performed on the reservoir coastal area with high terrain gradient, It will be reduced, thus avoiding the reservoir coastal areas with high gradient changes from being misjudged as deformation signals, reducing phase distortion, and performing phase unwrapping on areas with slight deformation. It will increase, strengthen the effect of slight deformation signals on phase unwrapping, and improve the ability to detect potential landslide disasters; when the pixel point belongs to a decoherent water area, the weight is set to 0, the unreliable phase of the area is excluded, and the path of the noise error of the water area to propagate to the global path is blocked, so as to avoid its noise phase from interfering with the global optimal unwrapping phase and causing distortion of the global deformation phase.

[0062] The objective function of the improved least-squares phase unwrapping algorithm, based on regional gradient adaptive regularization, adaptively adjusts the weight of the data fitting term based on the coherence and terrain gradient of different regions during phase unwrapping within the reservoir area. This ensures that the phase gradient in highly coherent regions dominates the optimization, while incoherent regions are excluded from the optimization, thus preventing random phase noise from influencing the global optimization process. In terrain transition zones, the terrain gradient of pixel points serves as a guide to improve the phase discrimination of the data fitting term for reservoir shorelines and potential landslide deformation areas. Furthermore, a gradient regularization term is introduced to eliminate the strong terrain gradients caused by the complex canyon terrain, preventing strongly varying terrain gradients from being misinterpreted as deformation signals. This further improves the accuracy of phase unwrapping in complex reservoir scenarios, enhances the quality of SAR image differential interferograms, and provides a reliable data foundation for subsequent landslide disaster detection.

[0063] Based on the constructed objective function, the optimization problem of the objective function is converted into solving the discrete Poisson equation to obtain the unwrapped phase according to the steps of the traditional least squares unwrapping algorithm. The discrete Poisson equation is solved using a method based on fast Fourier transform. The least squares unwrapping algorithm process based on fast Fourier transform is an existing technology, and its specific process will not be repeated here.

[0064] At this point, the phase unwrapping operation in the differential interferometry processing of the SAR image is completed, and the unwrapped interference phase is obtained. .

[0065] S5, based on the interference phase, obtain the deformation phase of the interference pattern, combine with the DEM data, and determine the elevation information corresponding to the interference pattern; according to the degree of change of the elevation information corresponding to the interference pattern at all time points, calculate the average deformation rate of each pixel point, and identify the potential landslide risk of the reservoir area corresponding to each pixel point.

[0066] Interference phase The deformation phase Terrain Phase , atmospheric phase and noise phase , where the terrain phase and atmospheric phase Estimated by external DEM data and atmospheric data, the noise phase Then, the interference phase Remove terrain phase , noise phase and atmospheric phase , the deformation phase of the interference pattern can be obtained .

[0067] According to the deformation phase , the DORIS phase-to-elevation method is used, combined with DEM to invert the elevation value according to the deformation phase, and a deformation interference map is obtained, which can reflect the elevation information of the deformation in the SAR image. The DORIS method is an existing technology, and its specific process will not be repeated here.

[0068] The number of SAR images collected in one year is recorded as T, then T-1 deformation interferograms can be obtained in one year, which are recorded as the multi-period deformation interferogram of the short baseline SAR data set.

[0069] After obtaining the multi-period deformation interferogram after differential interferometry processing, the PS-InSAR method is first used to perform time series analysis on the deformation interferogram to invert the cumulative deformation time series and calculate the annual average deformation rate of each pixel. , find out the potential landslide area. Arrange the annual average deformation rate of all pixels in descending order, and use the average of the previous preset percentage of annual average deformation rate as the judgment threshold , if there is an annual average deformation rate of the pixel point , then it is determined that the reservoir area corresponding to the pixel point has a landslide risk; otherwise, it is determined that there is no landslide risk. In this embodiment, the preset percentage is 10%, and the implementer can set it according to actual conditions. This embodiment does not impose any restrictions on this.

[0070] Furthermore, optical remote sensing data of the reservoir area was captured using drones equipped with optical remote sensing equipment. Geometric and atmospheric corrections were performed on the data to correct lens distortion, remove atmospheric interference with ground reflectivity, and ensure data consistency. Geometric and atmospheric corrections are currently available, and their specific processes are not detailed here.

[0071] The identification of potential landslide risks is verified by combining optical remote sensing data. Optical remote sensing data is used to focus on inspecting surface features such as steep bank slopes and drawdown zones in the reservoir area, as well as identified areas with landslide risks, to verify the accuracy of landslide risk area identification.

[0072] Based on the same inventive concept as the above method, an embodiment of the present application also provides a reservoir area landslide geological disaster identification system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-mentioned reservoir area landslide geological disaster identification methods.

[0073] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0074] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0075] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for identifying landslide geological hazards in reservoir areas, characterized in that: The method comprises the following steps: Collect SAR images and DEM data of the reservoir area at each time point; SAR images at adjacent time points are combined into a master and slave SAR image, and the interferogram of the master and slave SAR images is obtained. Based on the magnitude relationship between the coherence coefficients of the pixels of the master and slave SAR images, the interferogram of the master and slave SAR images is divided into a decoherent water area, a terrain transition area, and a high coherence stable area. The terrain phase of the reservoir area is determined by the geometric parameters of the DEM data and SAR image acquisition. The terrain phase gradient of each pixel point in the interference pattern is determined using the terrain phase, and the terrain gradient amplitude of each pixel point is obtained. The gradient amplitude of the unwrapped phase of each pixel in the interferogram is taken as an unknown term. According to the difference between the gradient amplitude of the entangled phase and the gradient amplitude of the unwrapped phase of each pixel in the interferogram, and the coherence coefficient deviation of the pixel points in the decoherent water area, the terrain transition area and the high coherence stable area, the difference is weighted. Combined with the terrain gradient amplitude, the objective function of the least squares phase unwrapping algorithm is modified to solve the unknown term and obtain the interference phase after phase unwrapping; Based on the interference phase, the deformation phase of the interference pattern is obtained, and the elevation information corresponding to the interference pattern is determined in combination with the DEM data; the average deformation rate of each pixel is calculated according to the degree of change of the elevation information corresponding to the interference pattern at all time points, and the potential landslide risk of the reservoir area corresponding to each pixel is identified; Using a multi-threshold segmentation algorithm, the coherence coefficients of all pixels in the interference pattern are segmented to obtain a first threshold and a second threshold, wherein the first threshold is less than the second threshold; When any pixel point in the interference map is located in the decoherent water area, the weight of the difference corresponding to the pixel point is 0; when any pixel point in the interference map is located in the terrain transition area, the mean of the first threshold and the second threshold is calculated, and the weight of the difference corresponding to the pixel point is the sum of the mean and the second threshold. The product of , where e is a natural constant, is the pixel point in the interference pattern The magnitude of the terrain gradient; is the pixel point in the interference pattern When any pixel point in the interference graph is located in a high coherence stable region, the weight of the any pixel point corresponding to the difference is the coherence coefficient of the any pixel point.

2. A method for identifying landslide geological hazards in a reservoir area according to claim 1, characterized in that: The interferogram of the master and slave SAR images is divided into a decoherent water area, a terrain transition area and a high coherence stable area, including: The area where the pixels with a coherence coefficient less than the first threshold in the interference map are located is regarded as the incoherent water area, the area where the pixels with a coherence coefficient greater than or equal to the first threshold and less than or equal to the second threshold in the interference map are located is regarded as the terrain transition area, and the area where the pixels with a coherence coefficient greater than the second threshold in the interference map are located is regarded as the high coherence stable area.

3. A method for identifying landslide geological hazards in a reservoir area according to claim 1, characterized in that: The topographic phase gradient includes a topographic phase gradient in the horizontal direction and a topographic phase gradient in the vertical direction of each pixel point.

4. A method for identifying landslide geological hazards in a reservoir area according to claim 2, characterized in that: The calculation formula for the terrain gradient amplitude is: Where, is the pixel point in the interference pattern The terrain phase gradient in the horizontal direction, is the pixel point in the interference pattern The terrain phase gradient in the vertical direction, is the horizontal and vertical coordinates of the pixel point.

5. A method for identifying landslide geological hazards in a reservoir area according to claim 1, characterized in that: The expression of the objective function is: Where, is the objective function of the interference pattern, is the pixel point in the interference pattern The weight corresponding to the area, is the size of the interference pattern, is the pixel point in the interference pattern The gradient amplitude of the unwrapped phase, is the pixel point in the interference pattern The gradient amplitude of the winding phase, is the preset smoothing factor, is the pixel point in the interference pattern The magnitude of the terrain gradient.

6. A method for identifying landslide geological hazards in a reservoir area according to claim 1, characterized in that: The obtaining of the deformation phase of the interference pattern includes: The terrain phase, the atmospheric phase and the noise phase in the interference phase after phase unwrapping are determined, and the terrain phase, the atmospheric phase and the noise phase are removed from the interference phase to obtain the deformation phase.

7. A method for identifying landslide geological hazards in a reservoir area according to claim 1, characterized in that: The identification of potential landslide risks in the reservoir area corresponding to each pixel point includes: The average deformation rates of all pixels are arranged in descending order, and the average of the first preset percentage of average deformation rates is used as the judgment threshold. The relationship between the average deformation rate of all pixels and the judgment threshold is used to identify the potential landslide risk in the reservoir area corresponding to each pixel.

8. A method for identifying landslide geological hazards in a reservoir area according to claim 7, characterized in that: The reservoir area corresponding to the pixels whose average deformation rate is greater than the judgment threshold is regarded as the landslide risk area.

9. A reservoir area landslide geological disaster identification system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images