Western plateau frozen soil hot melt disaster identification method based on remote sensing image

Abstract

The invention discloses a remote sensing image-based western plateau frozen soil hot melt disaster identification method. The method comprises the following steps of obtaining and preprocessing a satellite remote sensing image and a radar remote sensing image of a western plateau frozen soil preset area; performing multi-scale segmentation on the preprocessed satellite remote sensing image to obtain frozen soil characteristics of a preset area; performing time sequence InSA analysis on the preprocessed radar remote sensing image to obtain the earth surface deformation rate of the preset area; constructing a western plateau frozen soil hot melt disaster identification model based on the frozen soil characteristics and the earth surface deformation rate; and based on the western plateau frozen soil hot melt disaster identification model, completing western plateau frozen soil hot melt disaster identification based on the remote sensing image. According to the invention, the occurrence and evolution trend of frozen soil hot melt disasters can be found in time. By constructing a disaster recognition model, early warning and monitoring of potential disaster areas can be realized, and a scientific basis is provided for disaster prevention and control.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent identification of frozen soil thermal thaw disasters, and in particular relates to a method for identifying frozen soil thermal thaw disasters in the western plateau based on remote sensing images. Background Art

[0002] Western plateau regions, such as the Qinghai-Tibet Plateau, have vast permafrost areas with extreme climatic conditions, fragile ecological environments, and are sensitive to global climate change. Against the backdrop of global warming, permafrost areas in the western plateau are experiencing a strong degradation process, with frequent thermal melt disasters, such as thermal melt landslides and thermal melt lakes, posing a serious threat to the local ecological environment and engineering construction.

[0003] Remote sensing technology has the ability to quickly and macroscopically obtain data on the research area, can monitor surface changes in real time, and provides a rich data source for the identification of permafrost thermal thaw disasters. By integrating remote sensing data from different sensors, different temporal resolutions, and spatial resolutions, more comprehensive and accurate surface information can be obtained, improving the accuracy and reliability of disaster identification. With the development of computer technology and artificial intelligence, the automated processing and analysis capabilities of remote sensing images are constantly improving, which can achieve rapid processing and intelligent analysis of massive remote sensing data and improve the efficiency of disaster identification.

[0004] Although remote sensing technology has made significant progress in the identification of permafrost thermal thaw disasters, there are still some challenges and shortcomings. For example, the complexity and concealment of permafrost thermal thaw disasters make it difficult to identify; the permafrost characteristics and environmental conditions in different regions are different, requiring targeted identification methods; the accuracy and resolution of remote sensing data have an important impact on the identification results, etc. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a method for identifying thermal thaw disasters in the western plateau frozen soil based on remote sensing images, which can monitor the changes in the frozen soil surface in the western plateau in real time and give timely warning of thermal thaw disasters.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A method for identifying thermal thaw disasters of frozen soil in the western plateau based on remote sensing images comprises the following steps:

[0008] Acquire and pre-process satellite remote sensing images and radar remote sensing images of the preset permafrost areas of the western plateau;

[0009] Performing multi-scale segmentation on the pre-processed satellite remote sensing image to obtain frozen soil characteristics of a preset area;

[0010] Performing time series InSA analysis on the pre-processed radar remote sensing images to obtain the surface deformation rate of the preset area;

[0011] Based on the frozen soil characteristics and the surface deformation rate, a frozen soil thermal thaw disaster identification model for the western plateau is constructed;

[0012] Based on the western plateau permafrost thermal thaw disaster identification model, the western plateau permafrost thermal thaw disaster identification based on remote sensing images is completed.

[0013] Preferably, the method for preprocessing the satellite remote sensing image includes radiation correction, geometric correction and image enhancement; wherein the method for enhancing the satellite remote sensing image is:

[0014] Convert the satellite remote sensing image into HSV color space;

[0015] Performing gamma correction on the H and S spaces of the satellite remote sensing image to obtain a color-corrected image;

[0016] The V-space brightness of the satellite remote sensing image is enhanced by using a brightness curve to obtain a brightness enhanced image; wherein the brightness curve is symmetrically connected by seven convolutional layers;

[0017] Connecting the color-corrected image and the brightness-enhanced image in a channel direction to obtain a reconstructed image;

[0018] The reconstructed image is subjected to RGB space conversion to obtain a satellite remote sensing enhanced image.

[0019] Preferably, the method of performing multi-scale segmentation on the pre-processed satellite remote sensing image to obtain frozen soil characteristics is:

[0020] Preset multi-scale segmentation parameters, and perform a first segmentation on the satellite remote sensing enhanced image based on the preset multi-scale segmentation parameters, merge pixels to obtain a first polygon and the heterogeneity of the first polygon;

[0021] Performing a second segmentation based on the first polygon to obtain a second polygon and the heterogeneity of the second polygon;

[0022] When the obtained polygon heterogeneity meets the preset segmentation scale, the segmentation is terminated and a number of image segmentation objects are obtained;

[0023] Feature extraction is performed on each of the image segmentation objects to obtain the frozen soil features; wherein the frozen soil features include spectral features, contextual features, geometric features, spatial structure features, and texture features.

[0024] Preferably, the method for preprocessing the radar remote sensing image is:

[0025] Based on precise orbit parameters and a digital elevation model of a preset area, the radar remote sensing image is registered and cropped to obtain a registered radar remote sensing image;

[0026] Performing pairwise complex conjugate multiplication on the registered radar remote sensing images to obtain an interference image;

[0027] Adopting an adaptive window filtering algorithm to filter the interference image to obtain a filtered image;

[0028] Removing the flat ground phase portion of the interference phase information in the filtered image to obtain a horizontal interference image containing terrain changes;

[0029] Using a region growing algorithm, phase unwrapping is performed on the horizontal interference image to obtain an unwrapped phase between pixels in the interference image;

[0030] Reconstructing a digital elevation model based on the unwrapped phase, the imaging geometry of the registered radar remote sensing image, and orbital parameters;

[0031] Geographic coding is performed based on the digital elevation model, interference processing of the radar remote sensing image is completed, and an interference radar remote sensing image is obtained.

[0032] Preferably, the method of performing time series InSA analysis on the interferometric radar remote sensing images to obtain the surface deformation rate of the preset area is:

[0033] Performing amplitude analysis on the interferometric radar remote sensing image to obtain an initial pixel set including slow-loss correlation filter phase points;

[0034] Performing triangulation on the initial pixel set, and performing phase differentiation on adjacent points in the interferometric radar remote sensing image based on the triangulation to obtain differential phase;

[0035] The spatially de-correlated terrain phase error in the differential phase is removed, and when the phase difference value of adjacent points meets a preset threshold, the deformation phase is obtained;

[0036] Based on the deformation phase, construct a deformation phase matrix;

[0037] Solving the deformation phase matrix, obtaining the time deformation sequence and deformation average rate of the slow loss correlation filter phase point;

[0038] Based on the time deformation sequence and the average deformation rate, the surface deformation rate of the preset area is obtained.

[0039] Preferably, based on the frozen soil characteristics and the surface deformation rate, a method for constructing a frozen soil thermal thaw disaster identification model in the western plateau is:

[0040] Screening the frozen soil characteristics and the surface deformation rate to obtain key frozen soil characteristics and key surface deformation rate;

[0041] By using a linear discriminant method, the key frozen soil characteristics and the key surface deformation rate are integrated to obtain a comprehensive feature vector;

[0042] Based on the comprehensive feature vector, a convolutional neural network was used to construct a model for identifying thermal thaw disasters in the western plateau.

[0043] Preferably, the permafrost characteristics and the characteristics of the surface deformation rate that reflect the stability of permafrost, deformation trend and underground ice melting are screened.

[0044] Preferably, the method further includes predicting the occurrence time and disaster level of permafrost thermal thaw disaster in the western plateau based on the surface deformation rate.

[0045] Compared with the prior art, the invention has the following beneficial effects: the identification method based on remote sensing images can comprehensively utilize multiple characteristic information such as spectrum, shape, texture, etc., combined with data such as surface deformation rate, to achieve accurate identification of frozen soil thermal thaw disasters. Compared with traditional field investigation methods, the invention has higher identification accuracy and wider coverage. By constructing a disaster identification model, early warning and monitoring of potential disaster areas can be achieved, providing a scientific basis for disaster prevention and control. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in 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 paying creative labor.

[0047] Figure 1 This is a flow chart of a method for identifying thermal thaw disasters of frozen soil in the western plateau based on remote sensing images according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0050] Embodiment 1

[0051] like Figure 1 As shown, a method for identifying thermal thaw disasters of frozen soil in the western plateau based on remote sensing images includes the following steps:

[0052] S1: Acquire and preprocess satellite remote sensing images and radar remote sensing images of the preset permafrost areas in the western plateau;

[0053] Preprocessing satellite remote sensing images:

[0054] A further implementation method is that the method for preprocessing the satellite remote sensing image includes radiation correction, geometric correction and image enhancement; wherein the method for enhancing the satellite remote sensing image is:

[0055] Convert satellite remote sensing images into HSV color space;

[0056] Perform gamma correction on the H and S spaces of satellite remote sensing images to obtain color-corrected images;

[0057] The brightness curve is used to enhance the V-space brightness of the satellite remote sensing image to obtain a brightness enhanced image; wherein the brightness curve is symmetrically connected by seven convolutional layers, each layer consists of 32 convolutional kernels of size 3x3 and stride 1, and each convolutional layer is connected to an activation function.

[0058] In this embodiment, the designed brightness curve maps the brightness space of the weak-light image to the enhancement space, and provides a best-fitting curve for each pixel of the weak-light satellite remote sensing image to dynamically adjust the brightness.

[0059] Therefore, the transformation function corresponding to the brightness curve is as follows:

[0060] E n (x) = E n-1 (x)+A n E n-1 (x)(1-E n-1 (x)),

[0061] Among them, A n is the global pixel-level parameter matrix corresponding to the nth adjustment, E n is the image brightness after the nth adjustment, E n-1 is the image brightness after the n-1th adjustment, and x represents the pixel of the image.

[0062] Connect the color-corrected image and the brightness-enhanced image in the channel direction to obtain a reconstructed image;

[0063] The reconstructed image is converted into RGB space to obtain satellite remote sensing enhanced image.

[0064] Preprocessing radar remote sensing images:

[0065] A further implementation method is that the method for preprocessing the radar remote sensing image is:

[0066] Based on precise orbit parameters and the digital elevation model of the preset area, the radar remote sensing image is registered and cropped to obtain the registered radar remote sensing image;

[0067] Perform pairwise complex conjugate multiplication on the registered radar remote sensing images to obtain the interference image;

[0068] Adopting adaptive window filtering algorithm to filter the interference image and obtain the filtered image;

[0069] The flat ground phase part of the interferometric phase information in the filtered image is removed to obtain a horizontal interferometric image containing terrain changes;

[0070] The region growing algorithm is used to perform phase unwrapping on the horizontal interferometric image to obtain the unwrapped phase between pixels in the interferometric image.

[0071] Reconstruct the digital elevation model based on the imaging geometry of the unwrapped phase and registered radar remote sensing images and orbital parameters;

[0072] Based on the digital elevation model, geocoding is performed to complete the interference processing of radar remote sensing images and obtain interferometric radar remote sensing images.

[0073] S2: Perform multi-scale segmentation on the pre-processed satellite remote sensing images to obtain the frozen soil characteristics of the preset area;

[0074] A further implementation method is to perform multi-scale segmentation on the pre-processed satellite remote sensing image to obtain the frozen soil characteristics:

[0075] Preset multi-scale segmentation parameters, and perform a first segmentation on the satellite remote sensing enhanced image based on the preset multi-scale segmentation parameters, merge pixels to obtain a first polygon and the heterogeneity of the first polygon; in this embodiment, take a single pixel as the starting point, calculate the heterogeneity of the first polygon obtained by merging with adjacent pixels, and then perform heterogeneity judgment, if the preset threshold is met, the segmentation is terminated, if not, the next segmentation is performed.

[0076] A second segmentation is performed based on the first polygon to obtain a second polygon and the heterogeneity of the second polygon; the first polygon obtained by the first segmentation is used as a starting point, the heterogeneity of the second polygon obtained by merging with adjacent pixels is calculated, and the heterogeneity is judged.

[0077] When the obtained polygon heterogeneity meets the preset segmentation scale, the segmentation is terminated and a number of image segmentation objects are obtained;

[0078] Feature extraction is performed on each image segmentation object to obtain frozen soil features, wherein the frozen soil features include spectral features, contextual features, geometric features, spatial structure features and texture features.

[0079] In this embodiment, heterogeneity includes spectral heterogeneity and shape heterogeneity; wherein, spectral heterogeneity is obtained based on the number of layers, layer weights, and standard deviation of each layer of the satellite remote sensing enhanced image. Shape heterogeneity is described by a compactness factor and a smoothness factor. Shape heterogeneity is obtained based on the perimeter of the segmented region of the satellite remote sensing enhanced image and the perimeter and area of ​​the minimum circumscribed matrix. Based on the comprehensive measurement criteria, the weights of spectral heterogeneity and shape heterogeneity are obtained, and the heterogeneity index of the target object in the satellite remote sensing enhanced image is obtained.

[0080] In this embodiment, the mean variance is used to obtain the optimal multi-scale segmentation parameters from the spectral perspective. Specifically, after the satellite remote sensing enhanced image is segmented, the number of pure objects increases (i.e., images containing only frozen ground surfaces), and the spectral heterogeneity between them and adjacent objects increases, and the mean and variance of the objects at this scale also increase accordingly. Conversely, if the number of mixed objects increases after segmentation, the spectral differences between objects and the mean variance of the objects will become smaller. When the mean variance value reaches a maximum value, the spectral heterogeneity between objects is the largest, and the segmentation scale is the optimal scale at this time.

[0081] S3: Perform time series InSA analysis on the preprocessed radar remote sensing images to obtain the surface deformation rate of the preset area;

[0082] A further implementation method is to perform time series InSA analysis on interferometric radar remote sensing images to obtain the surface deformation rate of a preset area by:

[0083] The amplitude analysis of the interferometric radar remote sensing image is performed to obtain an initial pixel set containing the slow-loss correlation filter phase point; specifically, the amplitude deviation value is calculated according to the amplitude information of the interferometric radar remote sensing image to obtain the phase stability index, and the slow-loss correlation filter phase point is obtained based on the phase stability index. The winding phase of the slow-loss correlation filter phase point includes the phase change caused by the surface deformation in the radar line of sight direction, the phase change caused by the atmospheric delay, the phase change caused by the orbit error, the phase change caused by the elevation error, and the phase change caused by the noise.

[0084] Based on the phase error existing in the phase point of the slow-loss correlation filter, the entanglement phase unwrapping is performed. Specifically, the initial pixel set is triangulated based on an irregular triangulated network, and the phase difference of adjacent points in the interferometric radar remote sensing image is performed based on the triangulated network to obtain the differential phase;

[0085] The spatially decorrelated terrain phase error in the differential phase is removed, and the deformation phase is obtained when the phase difference value of adjacent points meets the preset threshold; that is, the phase change caused by the aforementioned surface deformation, atmospheric delay, orbit error, elevation error and the phase generated by noise are removed.

[0086] Based on the deformation phase, a deformation phase matrix is ​​constructed; the expression of the deformation phase matrix is ​​as follows:

[0087]

[0088] Among them, the matrix A M×N Each row corresponds to an interferometric radar remote sensing image, and each column corresponds to a radar remote sensing image. A vector representing the deformation phase of the radar remote sensing image at the Nth moment on the slow-loss correlation filter phase point to be determined; A vector representing the M unwrapped differential phases.

[0089] Solve the deformation phase matrix to obtain the time deformation sequence and average deformation rate of the slow loss correlation filter phase point;

[0090] Based on the time deformation sequence and the average deformation rate, the surface deformation rate of the preset area is obtained.

[0091] In this embodiment, in order to overcome the difference in deformation rate inversion results caused by the collective morphology of the plateau frozen ground surface and the radar line of sight angle, the annual average deformation rate of all acquired slow-loss correlation filter phase points should be projected to the same and the steepest slope direction of the slope. The speed projected to the maximum slope direction can be used to compare the deformation speeds of slopes in different directions, thereby further enriching the deformation speed information. Specifically, based on the speed measured by the radar along its line of sight and the proportional coefficient of the actual frozen ground surface three-dimensional displacement and the displacement measured by the slow-loss correlation filter phase point target. Among them, the proportional coefficient is obtained by the angle between the radar satellite line of sight and the optimal steep slope direction of the slope.

[0092] S4: Based on the characteristics of frozen soil and the rate of surface deformation, a model for identifying thermal thaw disasters in the western plateau is constructed;

[0093] A further implementation method is to construct a method for identifying the thermal thaw disaster of frozen soil in the western plateau based on the characteristics of frozen soil and the surface deformation rate:

[0094] Screening frozen soil characteristics and surface deformation rates to obtain key frozen soil characteristics and key surface deformation rates;

[0095] The linear discriminant method is used to integrate key frozen soil characteristics and key surface deformation rates to obtain a comprehensive feature vector.

[0096] Based on the comprehensive feature vector and the convolutional neural network, the construction of the identification model of thermal thaw disasters of frozen soil in the western plateau was completed.

[0097] In this embodiment, the convolutional neural network model includes an input layer, multiple convolutional layers, an activation layer, a pooling layer, a fully connected layer and an output layer.

[0098] According to the dimension and complexity of the comprehensive feature vector, adjust the number of convolutional layers, the size of the convolution kernel, the step size and other parameters.

[0099] After the convolutional layer, nonlinear activation functions such as ReLU are used to increase the nonlinear expression ability of the model.

[0100] The pooling layer is used to reduce the dimension of the feature map and reduce the amount of computation while retaining important features.

[0101] The fully connected layer converts the feature map output by the convolutional layer into a feature vector for classification or regression tasks.

[0102] A further implementation method is to screen the characteristics of frozen soil and the surface deformation rate that reflect the stability of frozen soil, deformation trend and underground ice melting.

[0103] S5: Based on the western plateau permafrost thermal thaw disaster identification model, complete the identification of permafrost thermal thaw disasters in the western plateau based on remote sensing images.

[0104] A further implementation method is to predict the occurrence time and disaster level of the frozen soil thermal thaw disaster in the western plateau based on the surface deformation rate. A frozen soil surface deformation rate curve is established, and the time node when the frozen soil thermal thaw disaster may occur is predicted by analyzing the surface deformation rate curve.

[0105] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for identifying thermal thaw disasters of frozen soil in the western plateau based on remote sensing images, characterized in that: The following steps are involved: Acquire and pre-process satellite remote sensing images and radar remote sensing images of the preset permafrost areas of the western plateau; Performing multi-scale segmentation on the pre-processed satellite remote sensing image to obtain frozen soil characteristics of a preset area; Performing time series InSA analysis on the pre-processed radar remote sensing images to obtain the surface deformation rate of the preset area; Based on the frozen soil characteristics and the surface deformation rate, a frozen soil thermal thaw disaster identification model for the western plateau is constructed; Based on the western plateau permafrost thermal thaw disaster identification model, the western plateau permafrost thermal thaw disaster identification based on remote sensing images is completed.

2. The method for identifying thermal thaw disasters of frozen soil in the western plateau based on remote sensing images according to claim 1 is characterized in that: The method for preprocessing the satellite remote sensing image includes radiation correction, geometric correction and image enhancement; wherein the method for enhancing the satellite remote sensing image is: Convert the satellite remote sensing image into HSV color space; Performing gamma correction on the H and S spaces of the satellite remote sensing image to obtain a color-corrected image; The V-space brightness of the satellite remote sensing image is enhanced by using a brightness curve to obtain a brightness enhanced image; wherein the brightness curve is symmetrically connected by seven convolutional layers; Connecting the color-corrected image and the brightness-enhanced image in a channel direction to obtain a reconstructed image; The reconstructed image is subjected to RGB space conversion to obtain a satellite remote sensing enhanced image.

3. The method for identifying thermal thaw disasters of frozen soil in the western plateau based on remote sensing images according to claim 2 is characterized in that: The method of performing multi-scale segmentation on the pre-processed satellite remote sensing image to obtain the frozen soil characteristics is as follows: Preset multi-scale segmentation parameters, and perform a first segmentation on the satellite remote sensing enhanced image based on the preset multi-scale segmentation parameters, merge pixels to obtain a first polygon and the heterogeneity of the first polygon; Performing a second segmentation based on the first polygon to obtain a second polygon and the heterogeneity of the second polygon; When the obtained polygon heterogeneity meets the preset segmentation scale, the segmentation is terminated and a number of image segmentation objects are obtained; Feature extraction is performed on each of the image segmentation objects to obtain the frozen soil features; wherein the frozen soil features include spectral features, contextual features, geometric features, spatial structure features, and texture features.

4. The method for identifying thermal thaw disasters of frozen soil in the western plateau based on remote sensing images according to claim 3 is characterized in that: The method for preprocessing the radar remote sensing image is: Based on precise orbit parameters and a digital elevation model of a preset area, the radar remote sensing image is registered and cropped to obtain a registered radar remote sensing image; Performing pairwise complex conjugate multiplication on the registered radar remote sensing images to obtain an interference image; Adopting an adaptive window filtering algorithm to filter the interference image to obtain a filtered image; Removing the flat ground phase portion of the interference phase information in the filtered image to obtain a horizontal interference image containing terrain changes; Using a region growing algorithm, phase unwrapping is performed on the horizontal interference image to obtain an unwrapped phase between pixels in the interference image; Reconstructing a digital elevation model based on the unwrapped phase, the imaging geometry of the registered radar remote sensing image, and orbital parameters; Geographic coding is performed based on the digital elevation model, interference processing of the radar remote sensing image is completed, and an interference radar remote sensing image is obtained.

5. The method for identifying thermal thaw disasters of frozen soil in the western plateau based on remote sensing images according to claim 4 is characterized in that: The method of performing time series InSA analysis on the interferometric radar remote sensing images to obtain the surface deformation rate of the preset area is as follows: Performing amplitude analysis on the interferometric radar remote sensing image to obtain an initial pixel set including slow-loss correlation filter phase points; Performing triangulation on the initial pixel set, and performing phase differentiation on adjacent points in the interferometric radar remote sensing image based on the triangulation to obtain differential phase; The spatially de-correlated terrain phase error in the differential phase is removed, and when the phase difference value of adjacent points meets a preset threshold, the deformation phase is obtained; Based on the deformation phase, construct a deformation phase matrix; Solving the deformation phase matrix, obtaining the time deformation sequence and deformation average rate of the slow loss correlation filter phase point; Based on the time deformation sequence and the average deformation rate, the surface deformation rate of the preset area is obtained.

6. The method for identifying thermal thaw disasters of frozen soil in the western plateau based on remote sensing images according to claim 1 is characterized in that: Based on the frozen soil characteristics and the surface deformation rate, the method for constructing the western plateau frozen soil thermal thaw disaster identification model is as follows: Screening the frozen soil characteristics and the surface deformation rate to obtain key frozen soil characteristics and key surface deformation rate; By using a linear discriminant method, the key frozen soil characteristics and the key surface deformation rate are integrated to obtain a comprehensive feature vector; Based on the comprehensive feature vector, a convolutional neural network was used to construct a model for identifying thermal thaw disasters in the western plateau.

7. The method for identifying thermal thaw disasters of frozen soil in the western plateau based on remote sensing images according to claim 6 is characterized in that: The frozen soil characteristics and the characteristics of the surface deformation rate that reflect the frozen soil stability, deformation trend and underground ice melting conditions are screened.

8. The method for identifying thermal thaw disasters of frozen soil in the western plateau based on remote sensing images according to claim 6 is characterized in that: It also includes predicting the occurrence time and disaster level of permafrost thaw disaster in the western plateau based on the surface deformation rate.

Images