A method for identifying ice cracks based on multi-source remote sensing data

Abstract

This invention relates to the field of image data processing, specifically to a method for identifying ice fissures based on multi-source remote sensing data. The method involves edge detection of the multi-source remote sensing data to obtain images of the ice fissure floor; preprocessing the ice fissure floor images to distinguish between glacier fissure development areas and non-glacier fissure areas; performing edge detection, tubular flow field identification, and dynamic threshold optimization on the glacier fissure development areas to extract glacier fissure data; and verifying the accuracy of the glacier fissure data to obtain the final glacier fissure distribution information. This invention achieves accurate ice fissure identification by improving the valley floor edge detection algorithm and dynamic threshold optimization strategy, combined with high-resolution imagery, effectively leveraging the advantages of remote sensing data to achieve low-cost identification and obtain more reliable data results.

Description

Technical Field

[0001] This invention relates to the field of image data processing, and in particular to a method for identifying ice cracks based on multi-source remote sensing data. Background Technology

[0002] Current research on ice crack extraction is limited, failing to meet the requirements for both adaptability and automated crack extraction. The mainstream method for ice crack extraction is visual interpretation, which allows for tracing back 50 years by manually identifying cracks in remote sensing images. Crack extraction sources include ground-penetrating radar (GPR), optical remote sensing images, synthetic aperture radar (SAR) images, and laser data. However, these methods generally suffer from high costs, low efficiency, and unsuitability for large-scale glacier crack monitoring.

[0003] High-precision optical satellite imagery can significantly improve the efficiency of ice crack identification. Histogram equalization processing of 10-meter resolution panchromatic images effectively enhances image contrast, enabling accurate identification of individual ice cracks. However, the above methods cannot dynamically assess the evolution of large-scale, disordered ice crack zones, and cannot accurately and efficiently determine the evolution of ice crack shape contours. Summary of the Invention

[0004] To address the challenges of identifying ice fissures due to the mixture of glacial shadows and ice-rock boundaries, and the limitations of existing methods relying solely on single-source remote sensing data in compensating for insufficient image coverage and limited data across different regions, thus hindering efficient and accurate identification of the dynamic evolution of ice fissures, this invention provides an ice fissure identification method based on multi-source remote sensing data. The method includes the following steps:

[0005] S100: Acquire multi-source remote sensing data; based on the multi-source remote sensing data, perform valley edge detection to obtain images of the bottom of ice fissures;

[0006] S200: Preprocess the image of the bottom of the ice fissure to distinguish between glacial fissure development areas and non-glacial fissure areas;

[0007] S300: Perform edge detection, tubular flow field identification, and dynamic threshold optimization on the glacier fissure development area to extract glacier fissure data;

[0008] S400: Verify the accuracy of the glacier fissure data to obtain the final glacier fissure distribution information.

[0009] Preferably, in S100, multi-source remote sensing data is acquired, specifically as follows:

[0010] The region is photographed by drones and satellites to obtain multi-source remote sensing data; wherein, the multi-source remote sensing data includes drone imagery and satellite remote sensing imagery.

[0011] Preferably, in S100, valley edge detection is performed based on the multi-source remote sensing data to obtain an image of the bottom of the ice fissure, specifically as follows:

[0012] The gray-level gradient direction of the multi-source remote sensing data is identified to obtain the lowest valley point of the region in the corresponding direction;

[0013] The locations of the lowest valleys in all directions of the region are smoothed to obtain images of the bottom of the ice fissures.

[0014] Preferably, in S200, before preprocessing the image of the bottom of the ice crevasse, the following steps are included:

[0015] The image at the bottom of the ice fissure is reduced by a preset ratio to obtain a reduced image;

[0016] The pixel grayscale of the reduced image is calculated based on the pixel grayscale of the image at the bottom of the ice fissure.

[0017] Preferably, in S200, the image of the bottom of the ice fissure is preprocessed to distinguish between glacial fissure development areas and non-glacial fissure areas, specifically as follows:

[0018] Crack segments are identified in the reduced image to determine the crack locations;

[0019] Based on the location of the cracks, glacial fissure development areas and non-glacial fissure areas can be distinguished.

[0020] Preferably, in S200, crack segment identification is performed on the reduced image to obtain the crack location, specifically as follows:

[0021] Pixel gray levels are obtained from the reduced image, and pixels belonging to crack segments and non-crack segments within the reduced image are distinguished based on the pixel gray levels, thereby obtaining the crack location.

[0022] Preferably, in S300, edge detection, tubular flow field identification, and dynamic threshold optimization are performed on the glacier fissure development area to extract glacier fissure data, specifically:

[0023] The Sobel operator is used to perform edge detection on the glacial crack development region to obtain the edge gradient field of the glacial crack development region; based on the edge gradient field, the image of the glacial crack development region is sharpened.

[0024] A tubular flow field analysis was performed on the glacier fissure development region to obtain the glacier fissure profile of the glacier fissure development region.

[0025] The glacier crack profile is subjected to dynamic threshold optimization processing to obtain glacier crack data in the glacier crack development area; wherein, the glacier crack data includes glacier crack profile evolution data.

[0026] Preferably, in S300, the Sobel operator is used to perform edge detection on the glacial crack development region to obtain the edge gradient field of the glacial crack development region; based on the edge gradient field, image sharpening processing is performed on the glacial crack development region, specifically as follows:

[0027] The Sobel operator was used to perform edge detection on the glacial crack development region to obtain the edge gradient fields of the glacial crack development region in the horizontal and vertical directions.

[0028] The gradient magnitude of the edge gradient field is standardized, and the Laplacian operator is used to sharpen the image of the glacial crack development area.

[0029] Preferably, in S300, a tubular flow field analysis is performed on the glacier crack development region to obtain the glacier crack profile of the glacier crack development region, specifically as follows:

[0030] A tubular flow field analysis was performed on the glacial crack development area to obtain the evolution process of the geometric activity profile of glacial crack propagation, thereby obtaining the profile in the main extension direction and crack width direction of the glacial crack.

[0031] In S300, dynamic threshold optimization is performed on the glacier crack profile to obtain glacier fissure data in the glacier crack development area, specifically as follows:

[0032] Dynamic threshold optimization was performed on the contours of the glacier fissures in the main extension direction and the fissure width direction to obtain the topological data of the glacier fissures in the glacier fissure development area.

[0033] Preferably, in step S400, the accuracy of the glacier fissure data is verified to obtain the final glacier fissure distribution information, specifically as follows:

[0034] A confusion matrix is ​​constructed by comparing expert-interpreted images, and the overall accuracy of the glacier fracture data is determined based on the confusion matrix.

[0035] The Kappa coefficient of the glacier fracture data is determined based on the overall accuracy, producer accuracy, and user accuracy.

[0036] Based on the Kappa coefficient, the reliability of the glacier fracture data is determined, and the reliable glacier fracture data is used as the final glacier fracture distribution information.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] First, this invention improves the valley edge detection algorithm and dynamic threshold optimization strategy, and combines high-resolution images to achieve accurate identification of ice cracks. The overall accuracy reaches 90.7% and the Kappa coefficient is 0.91, which is significantly better than the ice crack identification accuracy of the traditional threshold method.

[0039] Second, this invention is applicable to large-scale ice crack identification and can quickly and accurately identify the location of ice cracks. Considering the wide coverage and huge amount of information in high-resolution images, if the image size is too large, it is impossible to quickly identify whether there are ice cracks in the high-resolution image. By reducing the image to the minimum size, the ice crack information is preserved and some image noise is removed.

[0040] Third, this invention employs the Sobel operator for edge detection of ice crack images and the Laplacian operator as a second-order differential operator for image sharpening. Furthermore, based on the Sobel operator, it introduces an effective method for segmenting linear structures in digital images, segmenting based on the Tubular Flow Field (TuFF) through propagating geometrically active contours. Additionally, to improve the geometric smoothness of the segmentation results, a contour length regularization term is introduced into the numerical solution. After defining the crack indicator function, geometrically driven crack segmentation is achieved based on the level set theory framework. By defining an implicit function, the evolving contour is represented as the zero level set of the function. Its advantage lies in its implicit representation characteristics, enabling the evolving contour to adaptively change the topological structure of ice cracks, such as bifurcation and closure.

[0041] Fourth, this invention introduces a dynamic threshold optimization algorithm system to address the adaptive shortcomings of traditional thresholding methods. By constructing a convergent numerical optimization model, a gradient descent-based optimization strategy is adopted to intelligently solve the segmentation threshold. Specifically, candidate thresholds are first initialized based on the maximum inter-class variance criterion. Then, an objective function is established in the feature space to quantitatively evaluate the segmentation quality. The threshold parameters are continuously updated using the Jacobi iteration matrix until the convergence condition is met, thereby improving the segmentation robustness in complex ice environments. Furthermore, by improving the weight allocation mechanism of the classic mean iteration algorithm, a regional contrast factor is introduced to dynamically weight the grayscale mean in each iteration. At the same time, the candidate thresholds are regularized by combining the correlation of the neighborhood space, ensuring stable segmentation performance in both snow-covered and shadowed interference areas.

[0042] Fifth, this invention provides a more efficient and convenient method for extracting ice fissures, which is particularly suitable for scenes where the shadows of glaciers and ice-rock boundaries are mixed on steep mountain slopes. It can automatically and accurately identify fissures using remote sensing data, leveraging the advantages of remote sensing data, resulting in low identification costs and more reliable data results. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0044] Figure 1 This is a flowchart of an ice crack identification method based on multi-source remote sensing data provided by the present invention.

[0045] Figure 2 It is the process of detecting the edge of the valley.

[0046] Figure 3 This is a Sobel operator computation instance.

[0047] Figure 4 It is a crack extraction map obtained from edge detection and tubular flow field recognition.

[0048] Figure 5 This is a crack extraction image after dynamic threshold optimization.

[0049] Figure 6 This is a map showing the distribution of glacial fissures in a specific mountainous area.

[0050] Figure 7 This is a diagram showing the distribution pattern of glacial fissure length and direction in a specific mountainous area. Detailed Implementation

[0051] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for explaining the present invention and not for limiting the present invention. Furthermore, it should be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, not all structures. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of the present invention.

[0052] The terms "comprising" and "having," and any variations thereof, used in this invention are intended to cover non-exclusive inclusion. For example, a process, method, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus.

[0053] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0054] Please see Figure 1 As shown, this invention provides a method for identifying ice cracks based on multi-source remote sensing data. The method includes the following steps:

[0055] S100: Acquire multi-source remote sensing data; based on the multi-source remote sensing data, perform valley edge detection to obtain images of the bottom of ice fissures;

[0056] S200: Image preprocessing of the bottom of ice fissures to distinguish between glacial fissure development areas and non-glacial fissure areas;

[0057] S300: Edge detection, tubular flow field identification, and dynamic threshold optimization are performed on glacier fissure development areas to extract glacier fissure data;

[0058] S400: Verify the accuracy of glacier fissure data to obtain the final glacier fissure distribution information.

[0059] Furthermore, in S100, multi-source remote sensing data is acquired, specifically as follows:

[0060] The region was photographed by drones and satellites to obtain multi-source remote sensing data, which included drone images and satellite remote sensing images.

[0061] Considering the varying image capture accuracy and coverage of different remote sensing data sources, using a single remote sensing data source cannot obtain accurate images of the entire region. To compensate for the shortcomings of single remote sensing data sources in terms of insufficient coverage or limited image quantity for different regions, imagery is acquired through multiple methods, including drone and satellite imagery, to obtain multi-source remote sensing data for the glacier area. Specifically, drone aerial photography can be used to obtain imagery data of the glacier area behind the glacier lake, and / or visible light imagery data can be obtained from remote sensing satellites, thus forming multi-source remote sensing data. This multi-source remote sensing data includes, but is not limited to, drone imagery, satellite remote sensing imagery, and SAR imagery. Preferably, drone aerial photography can be used to obtain glacier cataloging data images of the glacier area; alternatively, GF-2 satellite, ASTER GDEM satellite, and Google Maps can be used to acquire GF-2 imagery, ASTER GDEM imagery, and Google imagery of the glacier location, thereby enriching the multi-source remote sensing data and providing sufficient and reliable image data sources for the identification of crevasses in the glacier area.

[0062] Furthermore, in S100, valley edge detection is performed based on multi-source remote sensing data to obtain images of the bottom of ice fissures, specifically:

[0063] Gray-level gradient direction identification is performed on multi-source remote sensing data to obtain the lowest valley point in the corresponding direction of the region;

[0064] The locations of the lowest valleys in all directions of the region are smoothed to obtain images of the bottom of the ice fissures.

[0065] Practical ice crack identification and extraction is affected by gradient grayscale variations and noise. To address this, this invention designs a dedicated edge detection algorithm for ice crack detection. (See also...) Figure 2 In an image of a mountain glacier, the overall lighting is uneven. The grayscale level of the valleys generally increases from bottom to top, but in some local areas, the grayscale level is lower due to cracks and shadows. To identify the direction of the grayscale gradient in the valleys, the lowest valley point in a certain direction within the region should be detected, and its direction should be marked as its location. For example... Figure 2 As shown, three lines, "1", "2", and "3", are marked on the left and right sides of the trapezoidal region, respectively. If the detected center pixel "0" is a valley candidate, its grayscale value must meet the following conditions: the grayscale value of the center pixel "0" should be lower than the grayscale value of line "1", the grayscale value of line "1" should be lower than the grayscale value of line "2", and the grayscale value of line "2" should be lower than the grayscale value of line "3". Next, a weighted average value for each line is calculated. During the weighting process, the center pixel should have a larger weight, while other pixels should have smaller weights. Since fractional derivatives can effectively smooth images, especially showing advantages in handling fine edges, fractional derivatives are used here to calculate the weight coefficients.

[0066] Valley points possess unique directional characteristics in four directions: front, back, left, and right. Therefore, detection should be performed separately in each of these four directions. Furthermore, multi-source remote sensing images contain a significant amount of noise. To reduce noise interference, a Gaussian smoothing function is used for filtering. This Gaussian smoothing function has a width parameter sigma (often referred to as the scale space parameter). The choice of parameter sigma depends on the distribution of white spots in the image. Those skilled in the art can set the value of parameter sigma based on the actual distribution of white spot sizes; this will not be discussed in detail here.

[0067] Furthermore, in S200, prior to image preprocessing of the ice fracture bottom, the following steps are included:

[0068] The image at the bottom of the ice fissure is reduced according to a preset ratio to obtain a reduced image;

[0069] Calculate the pixel grayscale of the reduced image based on the pixel grayscale of the image at the bottom of the ice crevasse.

[0070] Multi-source remote sensing data has high image resolution and wide coverage, resulting in a massive amount of information. Large image sizes make it difficult to quickly identify the presence of glacial fissures in high-resolution images. To efficiently distinguish between glacial fissure-developed areas and non-glacial fissure areas, the image needs to be reduced to its minimum size to preserve fissure information and remove some image noise. Specifically, the grayscale of the corresponding pixel in the reduced image can be calculated using the grayscale of four adjacent pixels in the original fissure bottom image. For example, let f(x, y) be the original fissure bottom image, where x and y represent the row and column of the original fissure bottom image, respectively, and x = 1, 2, 3...n, y = 1, 2, 3...m. The reduced image is then defined as f(x, y). k y k ), where x k =1、...、n / 2 k y k =1、...、m / 2 k k is a positive integer, and k≤K, n≥2 K m≥2 K By using the above method, the ice crack information of the original ice crack bottom image can be preserved to the maximum extent while reducing the size of the original ice crack bottom image, thereby improving the signal-to-noise ratio of the reduced image.

[0071] Furthermore, in S200, image preprocessing of the crevasse base distinguishes between glacial crevasse-developed areas and non-glacial crevasse areas, specifically:

[0072] Crack segments are identified in the reduced image to determine the crack location;

[0073] Based on the location of the cracks, regions with glacial fissure development and regions without glacial fissures can be distinguished.

[0074] Furthermore, in S200, crack segment identification is performed on the reduced image to obtain the crack location, specifically as follows:

[0075] Pixel gray levels are obtained from the reduced image. Based on the pixel gray levels, pixels in the reduced image that belong to the crack segment and those that do not belong to the crack segment are distinguished, thereby obtaining the crack location.

[0076] Furthermore, in S300, edge detection, tubular flow field identification, and dynamic threshold optimization are performed on the glacier fissure development area to extract glacier fissure data, specifically:

[0077] The Sobel operator is used to detect the edges of glacial crack development areas to obtain the edge gradient field of the glacial crack development areas; based on the edge gradient field, the image of the glacial crack development areas is sharpened.

[0078] Tubular flow field analysis was performed on the glacier fissure development area to obtain the glacier fissure profile of the glacier fissure development area.

[0079] Dynamic threshold optimization was performed on the glacier crack profile to obtain glacier fissure data in the glacier crack development area; the glacier fissure data included glacier crack profile evolution data.

[0080] After preprocessing the original image of the bottom of the glacial fissure to obtain a reduced image, a post-processing function based on fracture mechanics is used to identify whether a segment can represent part of a fissure. This identifies and calibrates the fissure location, facilitating further fissure processing and classification. This invention employs an edge detector to detect fissures in the reduced image. The edge detector algorithm assumes that the fissure gray level must be lower than that of the non-fissure area on the glacier surface. If the gray level is higher than that of the non-fissure area, an inverse operation is performed on the reduced image. If the reduced image contains both dark and bright fissures, fissure detection is performed separately on both the original reduced image and the inverse operation image to determine the fissure location. Based on the fissure location, glacial fissure development areas and non-glacial fissure areas are distinguished. Generally, areas covered by fissures are defined as glacial fissure development areas, and areas not covered by fissures are defined as non-glacial fissure areas.

[0081] Furthermore, in S300, the Sobel operator is used to perform edge detection on the glacial crack development region to obtain the edge gradient field of the glacial crack development region; based on the edge gradient field, image sharpening processing is performed on the glacial crack development region, specifically as follows:

[0082] The Sobel operator was used to detect the edge of the glacial crack development region, and the edge gradient fields of the glacial crack development region in the horizontal and vertical directions were obtained.

[0083] The gradient magnitude of the edge gradient field is standardized, and the Laplacian operator is used to sharpen the image of the glacier crack development area.

[0084] Please see Figure 3 This invention employs the Sobel operator for edge detection in glacial crevasse regions. The Sobel operator uses two convolutional kernels. and Edge detection is performed separately in the horizontal and vertical directions. Figure 3 Presentation using convolution kernels The horizontal edge detection process is performed. Specifically, the convolution kernel... and as follows:

[0085] .

[0086] Specifically, the following edge gradient field is first calculated using the Sobel operator described above:

[0087] Horizontal gradient Vertical gradient ;in, Images showing areas where glacial crevasses have developed;

[0088] Then, the gradient magnitude and direction at the edge are calculated using the horizontal and vertical gradients:

[0089] gradient magnitude ;

[0090] gradient direction ;

[0091] The gradient magnitude of the edge gradient field is standardized, and the Laplacian operator is used as a second-order differential operator to sharpen the image of the glacial crack development area.

[0092] Furthermore, in S300, tubular flow field analysis was performed on the glacier fissure development region to obtain the glacier fissure profile of the region, specifically:

[0093] Tubular flow field analysis was performed on the glacier crack development area to obtain the evolution process of the geometric activity profile of glacier crack propagation, thereby obtaining the profile in the main extension direction and crack width direction of the glacier crack.

[0094] By introducing the Tubular Flow Field (TuFF) method, an efficient method for segmenting linear structures in digital images, segmentation is performed by propagating geometric active contours, which are implemented using level sets. Geometric active contours Propagation under the influence of a tubular flow field is governed by the following equations:

[0095]

[0096] In the above governing equations, , This is the direction coefficient, used to control the influence on the evolution of the curve, and , This is a function indicating the direction of crack development; for locations with elongated structures... , The value is relatively high (generally around 1), especially for locations without elongated structures. , The value of is relatively low (usually 0); It is a geometric active profile The unit normal vector field at each point on the surface; , These represent the axial component and the normal component, respectively. Indicates the main extension direction of the crack. This indicates a direction orthogonal to the main crack propagation direction (e.g., the crack width direction). In the above governing equations, , The propagation rate of the profile along the main extension direction of the crack and along the crack width direction are controlled separately. Please refer to [link / reference]. Figure 4 (a) is the original map of the glacial crack development area, and (b) is the crack extraction map corresponding to the original map of the glacial crack development area. Furthermore, to improve the geometric smoothness of the crack segmentation results, a contour length regularization term is introduced into the numerical solution: ,in For smoothing coefficients, For geometric active contours divergence, For geometric active contours The integral segment, by feeding back the change in contour curvature, achieves edge smoothing and effectively suppresses contour distortion caused by noise.

[0097] In S300, dynamic threshold optimization is performed on the glacier crack profile to obtain glacier fissure data in the glacier crack development area, specifically:

[0098] Dynamic threshold optimization was performed on the contours of glacial fissures in the main extension direction and the width direction to obtain glacial fissure topology data in the glacial fissure development area.

[0099] To address the adaptive limitations of traditional thresholding methods, a dynamic threshold optimization algorithm is introduced. After defining a fracture development direction indicator function, geometrically driven fracture segmentation is achieved based on the level set theory framework. The level set theory framework defines an implicit function A... Geometric activity contour This is represented as the zero level set of the implicit function, i.e.: The advantage of the above method lies in its implicit representation properties, which allow the geometrically active profile to adaptively change topological structures such as crack bifurcation and closure. The evolution equation of the reconstructed level set is as follows:

[0100] .

[0101] The reconstructed level set evolution equations described above are numerically discretized by introducing an entropy-stable scheme and solved using the finite difference method, effectively addressing the numerical dissipation problem in the level set function evolution process. Specifically, the numerical dissipation is periodically maintained through reinitialization techniques. The distance function ensures the numerical stability of the evolutionary process. Please refer to [link / reference]. Figure 5 (a) shows the crack map before dynamic threshold optimization, and (b) shows the crack extraction map after dynamic threshold optimization. Furthermore, to improve the segmentation robustness of complex glacier surfaces, the weight allocation mechanism of the classic mean-based iterative algorithm was improved. A regional contrast factor was introduced to dynamically weight the gray-level mean in each iteration, and the candidate threshold was regularized by combining neighborhood spatial correlation, enabling the algorithm to maintain stable segmentation performance in both snow-covered and shadowed areas.

[0102] Furthermore, in S400, the accuracy of the glacier fissure data is verified to obtain the final glacier fissure distribution information, specifically:

[0103] A confusion matrix was constructed by comparing expert-interpreted images, and the overall accuracy of the glacier fracture data was determined based on the confusion matrix.

[0104] The Kappa coefficient of glacier fracture data is determined based on overall accuracy, producer accuracy, and user accuracy.

[0105] The reliability of glacier fracture data is determined based on the Kappa coefficient, and reliable glacier fracture data is used as the final information on glacier fracture distribution.

[0106] To quantitatively evaluate the performance of the glacial fracture data acquisition process, spatial overlay analysis was used for accuracy verification. The verification dataset consisted of 18 high-resolution remote sensing images of typical GF-2 glacial areas, encompassing different regions such as exposed ice areas, snow-covered areas, and moraine-covered areas. Fifteen fracture distribution areas with varying illumination, orientation, and size were selected, and a baseline ground truth map was established through expert visual interpretation to ensure pixel-level accuracy in fracture labeling.

[0107] Specifically, a confusion matrix is ​​constructed by comparing expert-interpreted images, and the overall accuracy (OA) of the glacier fracture data is determined based on the confusion matrix. The Kappa coefficient of the glacier fracture data is determined based on the overall accuracy, producer's accuracy (PA), and user's accuracy (UA). These accuracies reflect the accuracy of image classification from different perspectives, quantifying the false negative and false positive rates of the algorithm in fracture pixel identification. The Kappa coefficient (KC) is calculated using the corresponding formula to measure the consistency between the extraction results and the human interpretation results. The calculation formula is as follows: In the above formula, To measure the consistency ratio of observations, κ represents the expected consistency ratio. When κ is greater than a preset value (e.g., the preset value is 0.81), the reliability of the glacier fracture data is determined, and the reliable glacier fracture data is used as the final glacier fracture distribution information.

[0108] Specifically, the accuracy verification results compared with expert visual interpretation are shown in Table 1 below:

[0109] Table 1

[0110] method OA / % PA / % UA / % Kappa coefficient (κ) Traditional threshold segmentation method 76.2 68.5 72.1 0.62 U-Net benchmark model 85.4 81.3 79.6 0.78 Method of the present invention 0.98 0.95 0.87 0.91

[0111] As shown in Table 1 above, the present invention provides a more efficient and convenient method for extracting ice fissures, which is particularly suitable for scenes where the shadows of glaciers and ice-rock boundaries are mixed on steep mountain slopes. It can automatically and accurately identify fissures using remote sensing data, leveraging the advantages of remote sensing data, resulting in low identification costs and more reliable data results.

[0112] Please see Figures 6-7 The images show the distribution of fissures in the Jirepu Glacier in the Chongduipu Basin of the Boqu River in the central Himalayas, and the distribution pattern of glacier fissure length and direction. The method of this invention can comprehensively and accurately identify and extract the outline size distribution of glacier fissures, which will not be described in detail here.

[0113] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of a necessary general-purpose hardware platform, or by a combination of hardware and software. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a computer product. The present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0114] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Other embodiments may also be used. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying ice cracks based on multi-source remote sensing data, characterized in that, The method includes the following steps: S100: Acquire multi-source remote sensing data; based on the multi-source remote sensing data, perform valley edge detection to obtain images of the bottom of ice fissures; S200: Preprocess the image of the bottom of the ice fissure to distinguish between glacial fissure-developed areas and non-glacial fissure areas; wherein... Before preprocessing the image of the bottom of the ice crevice, the process includes: reducing the image of the bottom of the ice crevice by a preset ratio to obtain a reduced image; and calculating the pixel grayscale of the reduced image based on the pixel grayscale of the image of the bottom of the ice crevice. The images of the bottom of the ice fissures are preprocessed to distinguish between glacial fissure-developed areas and non-glacial fissure areas, specifically as follows: The crack location is obtained by identifying crack segments in the reduced image. Specifically, the pixel gray levels are obtained from the reduced image, and the pixels in the reduced image that belong to crack segments and those that do not belong to crack segments are distinguished based on the pixel gray levels, thereby obtaining the crack location. Based on the location of the cracks, glacial fissure development areas and non-glacial fissure areas can be distinguished. S300: Edge detection, tubular flow field identification, and dynamic threshold optimization are performed on the glacier fissure development area to extract glacier fissure data, specifically: The Sobel operator is used to perform edge detection on the glacier fissure development region to obtain the edge gradient field of the glacier fissure development region; based on the edge gradient field, the image of the glacier fissure development region is sharpened. A tubular flow field analysis was performed on the glacier fissure development region to obtain the glacier fissure profile of the region. The glacier crack profile is subjected to dynamic threshold optimization processing to obtain glacier crack data in the glacier crack development area; wherein, the glacier crack data includes glacier crack profile evolution data; S400: Verify the accuracy of the glacier fissure data to obtain the final glacier fissure distribution information.

2. The method according to claim 1, characterized in that, In S100, multi-source remote sensing data is acquired, specifically as follows: The region is photographed by drones and satellites to obtain multi-source remote sensing data; wherein, the multi-source remote sensing data includes drone imagery and satellite remote sensing imagery.

3. The method according to claim 2, characterized in that, In S100, based on the multi-source remote sensing data, valley edge detection is performed to obtain an image of the bottom of the ice fissure, specifically: The gray-level gradient direction of the multi-source remote sensing data is identified to obtain the lowest valley point of the region in the corresponding direction; The locations of the lowest valleys in all directions of the region are smoothed to obtain images of the bottom of the ice fissures.

4. The method according to claim 1, characterized in that, In S300, the Sobel operator is used to perform edge detection on the glacier fissure development region to obtain the edge gradient field of the glacier fissure development region; based on the edge gradient field, image sharpening processing is performed on the glacier fissure development region, specifically as follows: The Sobel operator is used to perform edge detection on the glacier fissure development region to obtain the edge gradient fields of the glacier fissure development region in the horizontal and vertical directions; The gradient magnitude of the edge gradient field is standardized, and the Laplacian operator is used to sharpen the image of the glacier fissure development region.

5. The method according to claim 4, characterized in that, In S300, a tubular flow field analysis is performed on the glacier fissure development region to obtain the glacier fissure profile of the region, specifically: A tubular flow field analysis was performed on the glacier fissure development area to obtain the evolution process of the geometric activity profile of glacier fissure propagation, thereby obtaining the profile in the main extension direction and the width direction of the glacier fissure. In S300, dynamic threshold optimization is performed on the glacier crack profile to obtain glacier crack data in the glacier crack development area, specifically: Dynamic threshold optimization was performed on the contours of the glacier fissures in the main extension direction and the fissure width direction to obtain the topological data of the glacier fissure development area.

6. The method according to claim 1, characterized in that, In S400, the accuracy of the glacier fissure data is verified to obtain the final glacier fissure distribution information, specifically: A confusion matrix is ​​constructed by comparing expert-interpreted images, and the overall accuracy of the glacier fracture data is determined based on the confusion matrix. The Kappa coefficient of the glacier fracture data is determined based on the overall accuracy, producer accuracy, and user accuracy. Based on the Kappa coefficient, the reliability of the glacier fracture data is determined, and the reliable glacier fracture data is used as the final glacier fracture distribution information.

Images