Geological disaster monitoring and early warning system and method

Abstract

The invention provides a geological disaster monitoring and early warning system and method, and the method comprises the steps: obtaining a disaster monitoring image through monitoring a ground surface region with cracks; determining a plurality of feature mark points based on shadow pixel points of the disaster monitoring image, performing morphological expansion on each feature mark point to obtain a plurality of expanded pixels, and determining a plurality of suspicious feature points based on all the expanded pixels; dividing a plurality of sub-image blocks according to all the suspicious feature points, determining image block centroids of the sub-image blocks, and obtaining shape descriptors based on the minimum bounding rectangles of the sub-image blocks and the image block centroids; and analyzing the disaster monitoring image according to the shape descriptors of the sub-image blocks and the evolution prediction model of the cracks, and outputting an analysis result to a geological disaster early warning center. By adopting the scheme of the invention, geological disaster analysis can be carried out in a complex illumination shadow environment, so that the accuracy of geological disaster monitoring and early warning is improved.

Description

Technical Field

[0001] The present application relates to the technical field of monitoring image analysis, and more specifically, to a geological disaster monitoring and early warning system and method. Background Art

[0002] Monitoring image analysis is a process of detecting, identifying, tracking and analyzing targets in images or videos based on computer vision and image processing technology. It is widely used in environmental monitoring, disaster warning, intelligent transportation, security monitoring and other fields. In addition, the combination of real-time processing and big data analysis has promoted the application of monitoring image analysis in large-scale monitoring systems. For example, in disaster monitoring, the combination of satellite images and ground sensor data can identify the risks of natural disasters such as fires and floods in real time and quickly make response plans. With the continuous optimization of algorithms, the accuracy, speed and application scope of monitoring image analysis will continue to expand and become the core supporting technology for various intelligent systems.

[0003] Geological disaster monitoring based on monitoring image analysis uses remote sensing technology and deep learning algorithms, combined with multi-source data, to monitor and assess the risks of geological disasters in real time. It obtains high-resolution image data through drone aerial photography and ground-based camera equipment, uses neural network models for target detection, and automatically identifies abnormal characteristics of geological disasters. At the same time, based on multi-source data fusion technology, it improves the accuracy and response speed of the monitoring system, providing key data support for disaster emergency response. However, in the existing technology, the target monitoring area is usually complex in terrain, with large changes in lighting, and there will be many shadow areas. These shadows will not only cover up the characteristics of the disaster (such as cracks, landslides, etc.), but may also produce similar visual manifestations to geological changes such as cracks and rock deformation, affecting the accurate identification of the target, making it difficult to distinguish the characteristics of the disaster, thereby reducing the accuracy of monitoring and early warning. Therefore, how to realize geological disaster analysis in a complex lighting and shadow environment, thereby improving the accuracy of geological disaster monitoring and early warning has become a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides a geological disaster monitoring and early warning system and method, which can realize geological disaster analysis in a complex lighting and shadow environment, thereby improving the accuracy of geological disaster monitoring and early warning.

[0005] In a first aspect, the present application provides a disaster monitoring image analysis method, the method comprising the following steps: Monitor surface areas with cracks and obtain disaster monitoring images; Determine multiple characteristic marking points of the disaster monitoring image based on the pixel gradient of each shadow pixel in the disaster monitoring image, perform morphological expansion on the pixel points adjacent to each characteristic marking point in the disaster monitoring image in combination with a preset morphological operator to obtain multiple expanded pixels, and then determine multiple suspicious characteristic points in the disaster monitoring image based on all the expanded pixels; Divide the disaster monitoring image into multiple sub-image blocks according to all suspicious feature points, select a sub-image block as a selected sub-image block, determine the image block centroid of the selected sub-image block, perform shape analysis on the selected sub-image block based on the minimum circumscribed rectangle of the selected sub-image block and the image block centroid, obtain a shape descriptor of the selected sub-image block, and continue to determine the shape descriptors of the remaining sub-image blocks; The disaster monitoring image is analyzed according to the shape descriptors of each sub-image block and the crack evolution prediction model, and the analysis results are output to the geological disaster early warning center.

[0006] In some embodiments, determining a plurality of characteristic marker points of the disaster monitoring image based on the pixel gradient of each shadow pixel point in the disaster monitoring image specifically includes: Performing shadow segmentation on the disaster monitoring image to obtain a plurality of shadow areas of the disaster monitoring image; Determine the pixel gradient corresponding to each shadow pixel in each shadow area; Based on the pixel gradient corresponding to each shadow pixel point in each shadow area combined with the pixel gradient threshold, multiple feature marker points in the disaster monitoring image are extracted from all the shadow pixels.

[0007] In some embodiments, the pixel points adjacent to each feature mark point in the disaster monitoring image are morphologically expanded in combination with a preset morphological operator to obtain a plurality of expanded pixels, specifically including: Setting a morphological operator of the disaster monitoring image; Based on the morphological operator, feature expansion is performed on pixel points adjacent to each feature marker point in the disaster monitoring image to obtain a plurality of expanded pixels.

[0008] In some embodiments, determining multiple suspicious feature points in the disaster monitoring image based on all extended pixels specifically includes: Acquiring the disaster monitoring image and the morphological operator; Traversing the disaster monitoring image by using the morphological operator to obtain a plurality of verification areas; A plurality of suspicious feature points in the disaster monitoring image are determined based on all verification areas and all expanded pixels.

[0009] In some embodiments, dividing the disaster monitoring image into a plurality of sub-image blocks according to all suspicious feature points specifically includes: Cluster all suspicious feature points to obtain multiple feature grouping sets; A sub-image block corresponding to each feature grouping set is constructed, thereby obtaining a plurality of sub-image blocks of the disaster monitoring image.

[0010] In some embodiments, analyzing the disaster monitoring image according to the shape descriptors of each sub-image block and the crack evolution prediction model, and outputting the analysis results to the geological disaster early warning center specifically includes: Determine a plurality of surface risk features based on shape descriptors of each sub-image block and a preset feature threshold; Extracting crack features of the disaster monitoring image from all surface risk features; Generate a disaster trend analysis diagram of the disaster monitoring image by combining the crack evolution prediction model with the crack characteristics; The disaster trend analysis diagram is used as the analysis result of the disaster monitoring image, and the analysis result is sent to a geological disaster early warning center.

[0011] In some embodiments, a drone is used to monitor the surface area where cracks exist and to capture disaster monitoring images.

[0012] In a second aspect, the present application provides a geological disaster monitoring and early warning system, which includes a disaster monitoring image analysis unit, characterized in that the disaster monitoring image analysis unit includes: An acquisition module is used to monitor the surface area where cracks exist and obtain disaster monitoring images; A processing module, used to determine a plurality of characteristic marking points of the disaster monitoring image based on the pixel gradient of each shadow pixel point in the disaster monitoring image, perform morphological expansion on the pixel points adjacent to each characteristic marking point in the disaster monitoring image in combination with a preset morphological operator to obtain a plurality of expanded pixels, and then determine a plurality of suspicious characteristic points in the disaster monitoring image based on all the expanded pixels; The processing module is further used to divide the disaster monitoring image into multiple sub-image blocks according to all suspicious feature points, select a sub-image block as a selected sub-image block, determine the image block centroid of the selected sub-image block, perform shape analysis on the selected sub-image block based on the minimum circumscribed rectangle of the selected sub-image block and the image block centroid, obtain a shape descriptor of the selected sub-image block, and continue to determine the shape descriptors of the remaining sub-image blocks; The execution module is used to analyze the disaster monitoring image according to the shape descriptors of each sub-image block and the crack evolution prediction model, and output the analysis results to the geological disaster early warning center.

[0013] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned disaster monitoring image analysis method.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned disaster monitoring image analysis method is implemented.

[0015] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects: In the geological disaster monitoring and early warning system and method provided in the present application, first, the surface area where cracks exist is monitored to obtain a disaster monitoring image; secondly, based on the pixel gradient of each shadow pixel point in the disaster monitoring image, multiple feature marking points of the disaster monitoring image are determined, and the pixel points adjacent to each feature marking point in the disaster monitoring image are morphologically extended in combination with a preset morphological operator to obtain multiple extended pixels, and then multiple suspicious feature points in the disaster monitoring image are determined based on all the extended pixels; further, multiple sub-image blocks of the disaster monitoring image are divided according to all the suspicious feature points, a sub-image block is selected as the selected sub-image block, the image block centroid of the selected sub-image block is determined, and the selected sub-image block is shape analyzed based on the minimum circumscribed rectangle of the selected sub-image block and the image block centroid to obtain the shape descriptor of the selected sub-image block, and the shape descriptors of the remaining sub-image blocks are further determined; finally, the disaster monitoring image is analyzed according to the shape descriptors of each sub-image block and the crack evolution prediction model, and the analysis results are output to the geological disaster early warning center.

[0016] It can be seen that the present application can realize geological disaster analysis in complex lighting and shadow environments, thereby improving the accuracy of geological disaster monitoring and early warning; first, obtaining disaster monitoring images can provide an effective data source for subsequent analysis and early warning; second, multiple feature marker points in the disaster monitoring image are determined to describe the approximate outline of the surface shadow area, thereby providing support for subsequent image analysis; further, all feature marker points are morphologically verified to obtain multiple suspicious feature points of the disaster monitoring image, so as to make a preliminary regional feature description of the surface cracks, thereby increasing the efficiency of recognition and processing in the subsequent analysis process; then, the disaster monitoring image is divided into multiple sub-image blocks, To extract the potential locations of geological disasters in surface monitoring images, thereby improving the accuracy of geological disaster monitoring; in addition, determining the shape descriptor of the sub-image block can provide data basis for subsequent judgment whether the sub-image block is a surface crack area, and can describe the geometric characteristics and expansion trend of the surface cracks, thereby improving the effectiveness of geological disaster monitoring; finally, the analysis results of the disaster monitoring image are output according to the shape descriptor, thereby assisting management personnel to assess the possibility of geological disasters and take corresponding emergency prevention measures; in summary, the technical solution provided by this application can realize geological disaster analysis in complex lighting and shadow environments, thereby improving the accuracy of geological disaster monitoring and early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is an exemplary flow chart of a disaster monitoring image analysis method according to some embodiments of the present application; Figure 2 is an exemplary flow chart of determining a plurality of characteristic marking points of the disaster monitoring image according to some embodiments of the present application; Figure 3 is an exemplary flow chart of determining a shape descriptor of a selected sub-image block according to some embodiments of the present application; Figure 4 is a schematic diagram of the structure of a disaster monitoring image analysis unit according to some embodiments of the present application; Figure 5 It is a structural schematic diagram of a computer device for implementing a disaster monitoring image analysis method according to some embodiments of the present application. DETAILED DESCRIPTION

[0018] In order to better understand the technical solution of the present application, the technical solution of the present application will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0019] refer to Figure 1 , which is an exemplary flow chart of a disaster monitoring image analysis method according to some embodiments of the present application. The disaster monitoring image analysis method 100 mainly includes the following steps: In step 101, the surface area where cracks exist is monitored to obtain a disaster monitoring image.

[0020] In a specific implementation, the surface area where cracks exist is monitored to obtain disaster monitoring images, that is, the surface area where cracks exist is monitored and photographed by a drone, and the images obtained are used as disaster monitoring images. In addition, in other embodiments, other methods can also be used to obtain disaster monitoring images, such as satellite images, which are not limited here.

[0021] It should be noted that the disaster monitoring images in this application refer to images used to monitor information related to natural disasters. The images have a wide coverage and high resolution. Obtaining disaster monitoring images can provide an effective data source for subsequent analysis and early warning.

[0022] In step 102, multiple feature marker points of the disaster monitoring image are determined based on the pixel gradient of each shadow pixel in the disaster monitoring image, and the pixel points adjacent to each feature marker point in the disaster monitoring image are morphologically expanded in combination with a preset morphological operator to obtain multiple expanded pixels, and then multiple suspicious feature points in the disaster monitoring image are determined based on all the expanded pixels.

[0023] In some embodiments, reference Figure 2 As shown in FIG. 1 , this figure is an exemplary flow chart of determining multiple characteristic marking points of the disaster monitoring image according to some embodiments of the present application. In this embodiment, determining multiple characteristic marking points of the disaster monitoring image based on the pixel gradient of each shadow pixel point in the disaster monitoring image can be implemented by the following steps: First, in step 1021, shadow segmentation is performed on the disaster monitoring image to obtain a plurality of shadow areas of the disaster monitoring image; Then, in step 1022, the pixel gradient corresponding to each shadow pixel point in each shadow area is determined; Finally, in step 1023, multiple feature marker points in the disaster monitoring image are extracted from all the shadow pixels based on the pixel gradient corresponding to each shadow pixel in each shadow area combined with the pixel gradient threshold.

[0024] In a specific implementation, the disaster monitoring image is shadow segmented by a threshold segmentation method to obtain multiple shadow areas of the disaster monitoring image. The threshold segmentation method is an existing technology for image processing. The specific segmentation process will not be repeated here. It should be noted that the shadow area represents the position of the black shadow generated in the area below the surface when light is irradiated. By segmenting the shadow area, the geological characteristics of the cracks can be further analyzed and extracted, thereby improving the effectiveness of the early warning.

[0025] In some embodiments, the pixel gradient corresponding to each shadow pixel point in each shadow area may be determined in the following manner, namely: For each shadow area, extract all shadow pixels in the shadow area; The pixel gradient corresponding to each shadow pixel is determined according to the grayscale value of each shadow pixel.

[0026] In a specific implementation, all shadow pixel points in the shadow area are extracted, that is, all pixel points located in the shadow area are taken as shadow pixel points, so as to obtain all shadow pixel points in the shadow area.

[0027] In a specific implementation, the pixel gradient corresponding to each shadow pixel is determined according to the grayscale value of each shadow pixel, that is: a shadow pixel is selected, and the horizontal pixel gradient and the vertical pixel gradient of the shadow pixel are obtained by combining the grayscale values ​​of the shadow pixel and its adjacent shadow pixels through the Canny edge operator, and the square value of the horizontal pixel gradient and the square value of the vertical pixel gradient are summed, and the arithmetic square root of the sum is used as the pixel gradient corresponding to the shadow pixel, thereby obtaining the pixel gradient corresponding to each shadow pixel. In addition, in other embodiments, other methods can be used to calculate the pixel gradient, such as the Soble edge operator, which is not limited here.

[0028] It should be noted that the Canny edge operator represents an existing technology for edge detection in image processing, and the specific calculation process will not be repeated here.

[0029] In a specific implementation, multiple feature marker points in the disaster monitoring image are extracted from all the shadow pixels based on the pixel gradients corresponding to each shadow pixel in each shadow area combined with a pixel gradient threshold, that is, a pixel gradient threshold is set, the pixel gradients of each shadow pixel are compared with the pixel gradient threshold respectively, all pixel gradients greater than the pixel gradient threshold are extracted, and the shadow pixel points corresponding to the extracted pixel gradients are used as feature marker points, thereby obtaining multiple feature marker points in the disaster monitoring image.

[0030] It should be noted that the value of the pixel gradient threshold in this embodiment can be set according to specific application requirements, and the specific size is not limited here. In addition, the feature marker points in this application represent edge pixel points located at the shadow position, which have relatively prominent color and light and dark characteristics. By extracting the feature marker points, the approximate outline of the shadow area on the surface can be described, thereby providing support for subsequent image analysis, thereby improving the accuracy of prediction of potential geological disasters.

[0031] In some embodiments, the pixel points adjacent to each feature mark point in the disaster monitoring image are morphologically expanded in combination with a preset morphological operator to obtain a plurality of expanded pixels in the following manner, namely: Setting a morphological operator of the disaster monitoring image; Based on the morphological operator, feature expansion is performed on pixel points adjacent to each feature marker point in the disaster monitoring image to obtain a plurality of expanded pixels.

[0032] In the specific implementation, the morphological operator of the disaster monitoring image is set, that is, a rectangular frame of a fixed size is set, and the rectangular frame is used as the morphological operator of the disaster monitoring image. The size of the morphological operator can be set according to the accuracy requirements in the actual application. There is no limitation here. The size set in this embodiment is 5*5.

[0033] It should be noted that the morphological operator in this embodiment represents an operator for further detecting and confirming the surface morphology of the disaster monitoring image. By setting the morphological operator, the robustness and consistency of image analysis and processing can be enhanced, thereby improving the accuracy of monitoring the disaster monitoring image.

[0034] In a specific implementation, feature expansion is performed on the pixel points adjacent to each feature marker point in the disaster monitoring image based on the morphological operator to obtain multiple extended pixels, that is: the morphological operator is used to traverse each feature marker point in turn starting from the upper left corner of the disaster monitoring image, and each traversal covers a pixel area, which contains the feature marker point and adjacent pixels within a range of 5*5 with the feature marker point as the center. If there are at least five other feature marker points in the pixel area, all the pixels in the pixel area are used as extended pixels, thereby obtaining multiple extended pixels of the disaster monitoring image.

[0035] It should be noted that in this embodiment, the expanded pixels represent the newly added pixel points in the process of repairing the integrity of the disaster monitoring image. By expanding the pixels, the shape and connectivity of the disaster monitoring image can be optimized, thereby enhancing the effectiveness of geological disaster monitoring.

[0036] In some embodiments, the following method may be used to determine multiple suspicious feature points in the disaster monitoring image based on all the extended pixels, namely: Acquiring the disaster monitoring image and the morphological operator; Traversing the disaster monitoring image by using the morphological operator to obtain a plurality of verification areas; A plurality of suspicious feature points in the disaster monitoring image are determined based on all verification areas and all expanded pixels.

[0037] In specific implementation, the disaster monitoring image is traversed by the morphological operator to obtain multiple verification areas, that is: starting from the upper left corner of the disaster monitoring image, the disaster monitoring image is traversed by the morphological operator with a sliding step size of 1, and the covered area of ​​each traversal is used as a verification area, thereby obtaining multiple verification areas.

[0038] It should be noted that, in this embodiment, the verification area refers to the sub-image area obtained by traversal, and all pixels in the area need to be further verified and analyzed.

[0039] In a specific implementation, multiple suspicious feature points in the disaster monitoring image are determined based on all verification areas and all extended pixels, that is, a verification area is selected, and if all pixels in the verification area are extended pixels, the central pixel of the verification area is used as a suspicious feature point, otherwise no processing is performed, thereby obtaining multiple suspicious feature points of the disaster monitoring image. It should be noted that the suspicious feature points in the present application represent pixel points suspected to be located at surface cracks. The suspicious feature points are identified in the process of geological disaster monitoring of disaster monitoring images. In the subsequent monitoring process, it is necessary to focus on analyzing and verifying the suspicious feature points. By extracting the suspicious feature points, a preliminary regional feature description of the location of the surface cracks can be made, thereby increasing the efficiency of identification and processing in the subsequent analysis process, and further improving the accuracy and timeliness of disaster monitoring and early warning.

[0040] In step 103, multiple sub-image blocks of the disaster monitoring image are divided according to all suspicious feature points, one sub-image block is selected as the selected sub-image block, the image block centroid of the selected sub-image block is determined, and a shape analysis is performed on the selected sub-image block based on the minimum circumscribed rectangle of the selected sub-image block and the image block centroid to obtain a shape descriptor of the selected sub-image block, and the shape descriptors of the remaining sub-image blocks are further determined.

[0041] In some embodiments, the following method may be used to divide the disaster monitoring image into multiple sub-image blocks according to all suspicious feature points, namely: Cluster all suspicious feature points to obtain multiple feature grouping sets; A sub-image block corresponding to each feature grouping set is constructed, thereby obtaining a plurality of sub-image blocks of the disaster monitoring image.

[0042] In the specific implementation, all suspicious feature points are clustered to obtain multiple feature grouping sets, that is: all suspicious feature points are clustered by a clustering algorithm to obtain multiple feature grouping sets. Specifically, the DBSCAN clustering algorithm can be used for clustering. The specific clustering process is not repeated here. In addition, in other embodiments, other algorithms can also be used for clustering, for example, K-means clustering, hierarchical clustering, etc., which are not limited here.

[0043] It should be noted that, in this embodiment, the feature grouping set represents a set of suspicious feature points, and determining the feature grouping set is helpful to improve the integrity of subsequent analysis.

[0044] In specific implementation, a sub-image block corresponding to each feature grouping set is constructed, that is, for each feature grouping set, all suspicious feature points in the feature grouping set are connected through a connectivity algorithm based on region growth to form a closed area, and the closed area is used as a sub-image block of the disaster monitoring image. The specific connection process is not repeated here. In addition, in other embodiments, other algorithms can also be used to connect pixel points in the feature grouping set, for example, pixel cluster merging based on machine learning, etc., which is not limited here.

[0045] It should be noted that the sub-image block in the present application represents the shadow area in the disaster monitoring image. The shadow area can represent potential surface cracks, which need to be further confirmed and analyzed. By determining the sub-image block, the location of potential geological disasters in the surface monitoring image can be extracted, thereby improving the accuracy of geological disaster monitoring.

[0046] In some embodiments, the image block centroid of the selected sub-image block may be determined in the following manner, namely: Extracting pixel coordinates of all pixels in the selected sub-image block; An image block centroid of the selected sub-image block is determined based on all pixel coordinates.

[0047] In specific implementation, the pixel coordinates of all pixels in the selected sub-image block can be extracted by image processing software Open CV, and the coordinate quantity of the pixel coordinates can be recorded. In addition, in other embodiments, other methods can be used to extract the pixel coordinates of the pixel points, which are not limited here.

[0048] In a specific implementation, the image block centroid of the selected sub-image block is determined based on all pixel coordinates, that is: the horizontal coordinate and the vertical coordinate corresponding to each pixel coordinate are extracted, and then all the horizontal coordinates and all the vertical coordinates are obtained, and the quotient of the cumulative sum of all the horizontal coordinates and the number of coordinates is taken as the centroid horizontal coordinate, and the quotient of the cumulative sum of all the vertical coordinates and the number of coordinates is taken as the centroid vertical coordinate, and the image block centroid of the selected sub-image block is determined by the centroid horizontal coordinate and the centroid vertical coordinate.

[0049] It should be noted that the image block centroid in the present application represents the geometric center of the sub-image block. By determining the image block centroid, subsequent geometric analysis of the sub-image block can be facilitated, thereby improving the accuracy of geological disaster monitoring.

[0050] In some embodiments, reference Figure 3 As shown in FIG. 1 , this figure is an exemplary flow chart of determining a shape descriptor of a selected sub-image block according to some embodiments of the present application. In this embodiment, a shape analysis is performed on the selected sub-image block based on the minimum bounding rectangle of the selected sub-image block and the centroid of the image block to obtain the shape descriptor of the selected sub-image block. The following steps can be used to implement the shape descriptor: First, in step 1031, the minimum bounding rectangle of the selected sub-image block is determined; Next, in step 1032, a first shape parameter of the selected sub-image block is determined based on the minimum bounding rectangle; Then, in step 1033, the maximum inscribed circle of the selected sub-image block is obtained by fitting based on the centroid of the image block; Then, in step 1034, a second shape parameter of the selected sub-image block is determined according to the maximum inscribed circle; Finally, in step 1035, a shape descriptor of the selected sub-image block is determined according to the first shape parameter and the second shape parameter.

[0051] In the specific implementation, the minimum bounding rectangle of the selected sub-image block is determined, that is, the pixel coordinates of all pixel points in the selected sub-image block are obtained, the maximum horizontal coordinate, the minimum horizontal coordinate, the maximum vertical coordinate, and the minimum vertical coordinate are extracted from all the pixel coordinates, and the minimum bounding rectangle of the selected sub-image block is determined by combining the maximum horizontal coordinate, the minimum horizontal coordinate, the maximum vertical coordinate, and the minimum vertical coordinate through the minAreaRect function in Open CV.

[0052] It should be noted that the minimum enclosing rectangle in the present application represents the smallest rectangle that can enclose the sub-image block, and the boundary of the rectangle is parallel to the coordinate axis. By determining the minimum enclosing rectangle, it is helpful to quickly evaluate the spatial range and geometric characteristics of the sub-image block, thereby simplifying the subsequent analysis process.

[0053] In a specific implementation, the first shape parameter of the selected sub-image block is determined based on the minimum circumscribed rectangle, that is, the diagonal length of the minimum circumscribed rectangle is used as the first shape parameter of the selected sub-image block. The first shape parameter represents the geometric characteristics of the sub-image block in shape, and provides data support for subsequent analysis.

[0054] In some embodiments, the maximum inscribed circle of the selected sub-image block can be obtained by fitting the centroid of the image block in the following manner, namely: Extract all boundary pixels of the selected sub-image block; Determining a boundary distance between each boundary pixel and the centroid of the image block; The maximum inscribed circle of the selected sub-image block is determined based on all boundary distances.

[0055] In a specific implementation, all boundary pixels of the selected sub-image block are extracted by a contour extraction algorithm, and the boundary pixels represent pixel points located at the boundary of the selected sub-image block. In addition, in other embodiments, other methods may be used to extract boundary pixels, such as connected domain marking, etc., which are not limited here.

[0056] In a specific implementation, the boundary distance between each boundary pixel and the centroid of the image block is determined, that is, a boundary pixel is selected, and the pixel coordinates of the boundary pixel and the pixel coordinates of the centroid of the image block are substituted into the distance calculation formula to obtain the boundary distance between the boundary pixel and the centroid of the image block, and then the boundary distance between each boundary pixel and the centroid of the image block is obtained.

[0057] It should be noted that, in this embodiment, the boundary distance refers to the distance from the image block centroid of the sub-image block to the boundary of the sub-image block.

[0058] In a specific implementation, the maximum inscribed circle of the selected sub-image block is determined based on all boundary distances, that is, the minimum boundary distance is extracted from all boundary distances, the boundary distance is used as the maximum inscribed radius, the center of mass of the image block is used as the circle center, and the maximum inscribed radius is used as the radius to obtain the maximum inscribed circle of the selected sub-image block.

[0059] It should be noted that, in this embodiment, the maximum inscribed circle represents a maximized circle that can be placed within the sub-image block and does not exceed the boundary of the sub-image block. By determining the maximum inscribed circle, more geometric features of the sub-image block can be extracted, thereby providing stronger data support for subsequent analysis.

[0060] In a specific implementation, the second shape parameter of the selected sub-image block is determined according to the maximum inscribed circle, that is, the pixel coordinates of all pixel points within the maximum inscribed circle are combined into a two-dimensional coordinate matrix, the covariance matrix of the two-dimensional coordinate matrix is ​​calculated, and the first eigenvalue and the second eigenvalue of the covariance matrix are calculated (the first eigenvalue is greater than the second eigenvalue), and the quotient of the first eigenvalue and the second eigenvalue is used as the second shape parameter of the selected sub-image block.

[0061] It should be noted that the above calculation process is implemented using the numpy library in Python. In addition, in other embodiments, other methods can be used for calculation, which are not limited here.

[0062] It should also be noted that, in this embodiment, the first eigenvalue and the second eigenvalue represent the degree of change of the coordinates of the pixel point in different directions. Therefore, in this embodiment, the second shape parameter represents the geometric characteristic parameter of the sub-image block in the direction.

[0063] In a specific implementation, a shape descriptor of the selected sub-image block is determined according to the first shape parameter and the second shape parameter, that is, the average of the first shape parameter and the second shape parameter is used as the shape descriptor of the selected sub-image block.

[0064] It should be noted that the shape descriptor in this application represents an indicator for measuring the geometric characteristics of a sub-image block. In geological disaster monitoring, there are often some isolated, small interference areas after preliminary extraction. By determining the shape descriptor, it can provide data basis for subsequent judgment of whether the sub-image block is a surface crack area, and can describe the geometric characteristics and expansion trend of the surface cracks, thereby improving the effectiveness of geological disaster monitoring.

[0065] It should be noted that the shape descriptors of the remaining sub-image blocks are further determined by the implementation step of "performing shape analysis on the selected sub-image block based on the minimum circumscribed rectangle of the selected sub-image block and the centroid of the image block to obtain the shape descriptor of the selected sub-image block", which will not be repeated here.

[0066] In step 104, the disaster monitoring image is analyzed according to the shape descriptors of each sub-image block and the crack evolution prediction model, and the analysis results are output to the geological disaster early warning center.

[0067] In some embodiments, the disaster monitoring image is analyzed according to the shape descriptors of each sub-image block and the crack evolution prediction model, and the analysis results are output to the geological disaster warning center in the following manner, namely: Determine a plurality of surface risk features based on shape descriptors of each sub-image block and a preset feature threshold; Extracting crack features of the disaster monitoring image from all surface risk features; Generate a disaster trend analysis diagram of the disaster monitoring image by combining the crack evolution prediction model with the crack characteristics; The disaster trend analysis diagram is used as the analysis result of the disaster monitoring image, and the analysis result is sent to a geological disaster early warning center.

[0068] In the specific implementation, multiple surface risk features are determined based on the shape descriptors of each sub-image block and a preset feature threshold, that is: a feature threshold is preset, the shape descriptors of all sub-image blocks are compared with the feature threshold, all shape descriptors greater than the feature threshold are extracted, and the sub-image blocks corresponding to the extracted shape descriptors are used as surface risk features, otherwise no processing is done, thereby obtaining multiple surface risk features.

[0069] It should be noted that the feature threshold in this embodiment represents a threshold for determining whether a sub-image block is risky. The specific value can be set based on actual monitoring needs and historical expert experience data. No limitation is made here. By setting the feature threshold, the interference caused by isolated and smaller interference areas to geological disaster monitoring can be eliminated.

[0070] It should also be noted that the surface risk characteristics in this embodiment represent the location of cracks on the surface. The existence of such locations will bring potential geological disaster risks and require special attention.

[0071] In specific implementation, the crack features of the disaster monitoring image are extracted from all surface risk features, that is, the shape descriptors of all surface risk features are obtained, the largest shape descriptor is extracted, and the surface risk features corresponding to the extracted shape descriptors are used as the crack features of the disaster monitoring image.

[0072] It should be noted that the crack characteristics in this application represent the most representative crack locations on the surface, which are often located in special parts of the disaster monitoring image, such as the upper or lower edge of the slope. By determining the crack characteristics, the most risky cracks in the disaster monitoring image can be characterized, so that they can be monitored in a focused manner, thereby improving the effectiveness of geological disaster monitoring and early warning.

[0073] In specific implementation, a disaster trend analysis diagram of the disaster monitoring image is generated by combining the crack characteristics with the crack evolution prediction model, that is: first, the crack characteristics are continuously photographed by a drone to obtain crack inspection data, and the shooting frequency can be set according to actual emergency needs, for example, shooting every ten minutes; then, the crack inspection data is input into the pre-trained crack evolution prediction model to obtain the change information of the crack characteristics; finally, the change information of the crack characteristics over time is integrated into a time series diagram to obtain the disaster trend analysis diagram of the disaster monitoring image.

[0074] It should be noted that the crack evolution prediction model in this embodiment is a neural network model trained by a large amount of disaster monitoring data. In addition, the information on the changes in crack characteristics over time can be integrated through GIS (Geographic Information System, GIS). In addition, in other embodiments, other methods can be used for information integration, which are not limited here.

[0075] It should also be noted that the crack inspection data in this embodiment represents a set of images obtained by a drone shooting the crack area at a fixed frequency. The disaster trend analysis chart contains change information such as the development trend and expansion rate of the crack characteristics. By generating a disaster trend analysis chart, management personnel can be assisted in assessing the possibility of geological disasters and taking corresponding emergency prevention measures.

[0076] In addition, in another aspect of the present application, in some embodiments, the present application provides a geological disaster monitoring and early warning system, the system includes a disaster monitoring image analysis unit, referring to Figure 4 , which is a schematic diagram of the structure of a disaster monitoring image analysis unit according to some embodiments of the present application, the disaster monitoring image analysis unit 200 includes: an acquisition module 201, a processing module 202 and an execution module 203, which are respectively described as follows: Acquisition module 201, in this application, acquisition module 201 is mainly used to monitor the surface area where cracks exist and acquire disaster monitoring images; Processing module 202, in the present application, the processing module 202 is mainly used to determine multiple feature marker points of the disaster monitoring image based on the pixel gradient of each shadow pixel point in the disaster monitoring image, and perform morphological expansion on the pixel points adjacent to each feature marker point in the disaster monitoring image in combination with a preset morphological operator to obtain multiple expanded pixels, and then determine multiple suspicious feature points in the disaster monitoring image based on all the expanded pixels; The processing module 202 is further used to divide the disaster monitoring image into multiple sub-image blocks according to all suspicious feature points, select a sub-image block as a selected sub-image block, determine the image block centroid of the selected sub-image block, perform shape analysis on the selected sub-image block based on the minimum circumscribed rectangle of the selected sub-image block and the image block centroid to obtain a shape descriptor of the selected sub-image block, and continue to determine the shape descriptors of the remaining sub-image blocks; Execution module 203, in the present application, the execution module 203 is mainly used to analyze the disaster monitoring image according to the shape descriptors of each sub-image block and the evolution prediction model of the crack, and output the analysis results to the geological disaster warning center.

[0077] In addition, the present application also provides a computer device, which includes a memory and a processor, the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned disaster monitoring image analysis method.

[0078] In some embodiments, reference Figure 5 , which is a schematic diagram of the structure of a computer device for implementing a disaster monitoring image analysis method according to some embodiments of the present application. The disaster monitoring image analysis method in the above embodiment can be Figure 5 The computer device 300 shown in the figure is implemented, and the computer device 300 includes at least one processor 301, a communication bus 302, a memory 303 and at least one communication interface 304.

[0079] The processor 301 may be a general-purpose central processing unit (CPU), or an application-specific integrated circuit (ASIC) or one or more processors for controlling the execution of the disaster monitoring image analysis method in the present application.

[0080] The communication bus 302 may be used to transmit information between the above-mentioned components.

[0081] The memory 303 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 303 may exist independently and be connected to the processor 301 via the communication bus 302. The memory 303 may also be integrated with the processor 301.

[0082] The memory 303 is used to store the program code for executing the solution of the present application, and the execution is controlled by the processor 301. The processor 301 is used to execute the program code stored in the memory 303. The program code may include one or more software modules. The determination of the disaster monitoring image analysis method in the above embodiment can be implemented by the processor 301 and one or more software modules in the program code in the memory 303.

[0083] The communication interface 304 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0084] In a specific implementation, as an embodiment, a computer device may include multiple processors, each of which may be a single-CPU processor or a multi-CPU processor. The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0085] The above-mentioned computer device may be a general-purpose computer device or a special-purpose computer device. In a specific implementation, the computer device may be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device or an embedded device. The embodiment of the present application does not limit the type of computer device.

[0086] In addition, the present application also provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the above-mentioned disaster monitoring image analysis method is implemented.

[0087] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0088] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A disaster monitoring image analysis method, characterized in that: The steps include: Monitor surface areas with cracks and obtain disaster monitoring images; Determine multiple characteristic marking points of the disaster monitoring image based on the pixel gradient of each shadow pixel in the disaster monitoring image, perform morphological expansion on the pixel points adjacent to each characteristic marking point in the disaster monitoring image in combination with a preset morphological operator to obtain multiple expanded pixels, and then determine multiple suspicious characteristic points in the disaster monitoring image based on all the expanded pixels; Divide the disaster monitoring image into multiple sub-image blocks according to all suspicious feature points, select a sub-image block as a selected sub-image block, determine the image block centroid of the selected sub-image block, perform shape analysis on the selected sub-image block based on the minimum circumscribed rectangle of the selected sub-image block and the image block centroid, obtain a shape descriptor of the selected sub-image block, and continue to determine the shape descriptors of the remaining sub-image blocks; The disaster monitoring image is analyzed according to the shape descriptors of each sub-image block and the evolution prediction model of the crack, and the analysis result is output to the geological disaster early warning center.

2. The method according to claim 1, characterized in that Determining a plurality of characteristic marking points of the disaster monitoring image based on the pixel gradient of each shadow pixel point in the disaster monitoring image specifically includes: Performing shadow segmentation on the disaster monitoring image to obtain a plurality of shadow areas of the disaster monitoring image; Determine the pixel gradient corresponding to each shadow pixel in each shadow area; Based on the pixel gradient corresponding to each shadow pixel point in each shadow area combined with the pixel gradient threshold, multiple feature marker points in the disaster monitoring image are extracted from all the shadow pixels.

3. The method according to claim 1, characterized in that Combining a preset morphological operator to perform morphological expansion on the pixel points adjacent to each feature mark point in the disaster monitoring image, the obtained multiple expanded pixels specifically include: Setting a morphological operator of the disaster monitoring image; Based on the morphological operator, feature expansion is performed on pixel points adjacent to each feature marker point in the disaster monitoring image to obtain a plurality of expanded pixels.

4. The method according to claim 1, characterized in that Determining multiple suspicious feature points in the disaster monitoring image based on all the extended pixels specifically includes: Acquiring the disaster monitoring image and the morphological operator; Traversing the disaster monitoring image by using the morphological operator to obtain a plurality of verification areas; A plurality of suspicious feature points in the disaster monitoring image are determined based on all verification areas and all expanded pixels.

5. The method according to claim 1, characterized in that Dividing the disaster monitoring image into a plurality of sub-image blocks according to all suspicious feature points specifically includes: Cluster all suspicious feature points to obtain multiple feature grouping sets; A sub-image block corresponding to each feature grouping set is constructed, thereby obtaining a plurality of sub-image blocks of the disaster monitoring image.

6. The method according to claim 1, characterized in that The disaster monitoring image is analyzed according to the shape descriptors of each sub-image block and the crack evolution prediction model, and the analysis results are output to the geological disaster early warning center, which specifically includes: Determine a plurality of surface risk features based on shape descriptors of each sub-image block and a preset feature threshold; Extracting crack features of the disaster monitoring image from all surface risk features; Generate a disaster trend analysis diagram of the disaster monitoring image by combining the crack evolution prediction model with the crack characteristics; The disaster trend analysis diagram is used as the analysis result of the disaster monitoring image, and the analysis result is sent to a geological disaster early warning center.

7. The method according to claim 1, characterized in that Use drones to monitor surface areas where cracks exist and take photos to obtain disaster monitoring images.

8. A geological disaster monitoring and early warning system, the geological disaster monitoring and early warning system includes a disaster monitoring image analysis unit, characterized in that: The disaster monitoring image analysis unit comprises: An acquisition module is used to monitor the surface area where cracks exist and obtain disaster monitoring images; A processing module, used to determine a plurality of characteristic marking points of the disaster monitoring image based on the pixel gradient of each shadow pixel point in the disaster monitoring image, perform morphological expansion on the pixel points adjacent to each characteristic marking point in the disaster monitoring image in combination with a preset morphological operator to obtain a plurality of expanded pixels, and then determine a plurality of suspicious characteristic points in the disaster monitoring image based on all the expanded pixels; The processing module is further used to divide the disaster monitoring image into multiple sub-image blocks according to all suspicious feature points, select a sub-image block as a selected sub-image block, determine the image block centroid of the selected sub-image block, perform shape analysis on the selected sub-image block based on the minimum circumscribed rectangle of the selected sub-image block and the image block centroid, obtain a shape descriptor of the selected sub-image block, and continue to determine the shape descriptors of the remaining sub-image blocks; The execution module is used to analyze the disaster monitoring image according to the shape descriptors of each sub-image block and the crack evolution prediction model, and output the analysis results to the geological disaster early warning center.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores codes, and the processor is configured to obtain the codes and execute the disaster monitoring image analysis method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the disaster monitoring image analysis method according to any one of claims 1 to 7 is implemented.

Images