Tunnel apparent disease detection method

Abstract

The invention discloses a tunnel apparent disease detection method. The method comprises the following steps: collecting three-dimensional point cloud data in a tunnel through a moving vehicle carrying a three-dimensional laser scanner; calculating a depth value according to a Z value of the three-dimensional point cloud data, and mapping the depth value to a gray scale range of 0-255 to obtain a gray scale image; identifying a gray value abrupt change area in the gray image through an edge detection algorithm to obtain an abrupt change crack area on the surface of the tunnel lining; performing gray segmentation on the gray value mutation region to obtain a water leakage region, and performing connected domain analysis on the water leakage region to obtain the area of the water leakage region; identifying a gray value mutation region according to a region growing algorithm to obtain a lining surface layer layering region; according to a template matching method, a drainage ditch area in the gray level image is identified; and marking tunnel cracks, water leakage, lining surface layer peeling and drainage ditch damage areas. According to the invention, the three-dimensional laser scanning and image processing technology is adopted to detect the apparent diseases of the tunnel, so that the disease maintenance efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel disease detection, and in particular to a method for detecting tunnel apparent diseases. Background Art

[0002] Tunnel surface defects mainly refer to various damage phenomena on the surface of the tunnel lining, such as cracks, water leakage, peeling, delamination, exposed steel bars, etc. These defects not only affect the appearance of the tunnel, but more importantly, they may endanger the safety and stability of the tunnel structure and even lead to major repairs or reconstruction of the tunnel.

[0003] The existing methods for detecting tunnel surface defects mainly include visual inspection, micro-damage detection, and ultrasonic detection. Figure 2 (As shown) Through manual visual inspection of the inner wall of the tunnel, cracks, water leakage and other defects are identified. It is greatly affected by human factors, the detection effect is poor, it is difficult to find hidden defects, and it is time-consuming and labor-intensive. Micro-destructive detection methods include coring and pull-out methods, which are used to detect the strength of concrete linings. The coring method directly observes the internal structure of concrete by drilling and sampling, and determines its strength and whether there are defects. The pull-out method uses a pull-out instrument to detect the pull-out force of anchors in solid concrete to determine the compressive strength of concrete. The coring method is time-consuming and inefficient, and it is easy to damage the tunnel waterproofing layer, affecting the appearance of the tunnel; the pull-out method requires professional equipment and technicians to operate, and the cost is relatively high. The ultrasonic detection method uses ultrasonic pulses to penetrate concrete and judge whether there are defects inside the concrete based on the reflected waves. The defect is that there is still a lack of detection accuracy, and it is difficult to accurately judge the severity and scope of the disease.

[0004] In order to solve the above problems, the present invention proposes a method for detecting tunnel surface defects, which realizes the detection of tunnel surface defects through three-dimensional laser scanning and image processing technology, so as to improve the efficiency of defect detection and maintenance. Summary of the invention

[0005] In view of the above problems, the present invention provides a method for detecting apparent defects in a tunnel, comprising: Collecting 3D point cloud data of the interior of the tunnel by a mobile vehicle equipped with a 3D laser scanner, wherein the 3D point cloud data includes tunnel cracks, water leakage, lining surface delamination, and drainage ditch damage areas; Calculate the depth value according to the Z value of the three-dimensional point cloud data, and map it to a grayscale range of 0-255 to obtain a grayscale image; The grayscale mutation region in the grayscale image is identified by an edge detection algorithm to obtain the mutation crack region on the tunnel lining surface; the grayscale mutation region is segmented to obtain the leakage region, and the connected domain analysis is performed on the seepage region to obtain the area of ​​the leakage region; the grayscale mutation region is identified by a region growing algorithm to obtain the lining surface layer region; the drainage ditch region in the grayscale image is segmented by a template matching method; In the original image corresponding to the three-dimensional point cloud data, the positions and areas of the tunnel cracks, water leakage, lining surface peeling and drainage ditch damage areas are marked.

[0006] In an optional manner, the step of identifying the grayscale value mutation region in the grayscale image by using an edge detection algorithm to obtain the mutation crack region on the tunnel lining surface further comprises: Smoothing the grayscale image using a Gaussian filter, and calculating the gradient amplitude of each pixel in the grayscale image using a first-order differential operator, wherein the gradient amplitude is the rate of change of the grayscale value at the pixel; Determine the pixel points whose gradient amplitude exceeds the preset high threshold as strong edge points, determine the pixel points between the preset low threshold and the preset high threshold as weak edge points, and determine the pixel points below the preset low threshold as non-edge points; The weak edge points are connected to the strong edge points to form an edge profile, thereby obtaining a sudden crack area on the tunnel lining surface.

[0007] In an optional manner, the calculation formula for the gradient amplitude is:

[0008] in, is the Sobel operator in the horizontal direction, ; is the Sobel operator in the vertical direction, ; is the pixel horizontal coordinate; is the pixel vertical coordinate.

[0009] In an optional manner, performing grayscale segmentation on the grayscale value mutation area to obtain the water leakage area further includes: Initialize the weights and means of the foreground and background in the gray value mutation area according to the Otsu algorithm; The inter-class variance is calculated according to the weight and the mean, and the gray value mutation area is segmented into a water leakage area according to the maximum threshold of the inter-class variance.

[0010] In an optional manner, the calculation formula for the between-class variance is:

[0011] in, , Grayscale The ratio of the number of pixels in the image to the total number of pixels in the image, is the preset threshold; ; ; , is the number of gray levels; , is the mean value of the entire image.

[0012] In an optional manner, the performing a connected domain analysis on the infiltrated water area to obtain the area of ​​the leaked water area further comprises: Traversing the binary image in the infiltration water area by depth-first search (DFS), marking interconnected white pixels as the same connected domain; The number of pixels contained in each marked connected domain is calculated to obtain the area of ​​the water leakage region.

[0013] In an optional manner, the step of identifying the gray value mutation region according to the region growing algorithm to obtain the lining surface delamination region further comprises: Starting from the initial seed point of the gray value mutation area, searching for its neighboring pixels; Determine whether the neighborhood pixel meets the adding condition according to the growth criterion, if so, add it to the current area and regard it as a new seed point; During the region growing process, the gray value difference of each pixel when it is added to the region is recorded. When the gray value difference of each pixel exceeds a preset threshold, the pixel is located in the gray value mutation region; The lining surface delamination area is determined according to the gray value mutation area.

[0014] In an optional manner, the separating the drainage ditch area in the grayscale image according to the template matching method further includes: Presetting a drainage ditch area as a template, wherein the template includes the shape, size and grayscale distribution of the drainage ditch; An area similar to the template is searched in the grayscale image, and an area whose similarity exceeds a preset threshold is determined as a drainage ditch area.

[0015] In an optional manner, after marking the locations and areas of the tunnel cracks, water leakage, lining surface delamination, and drainage ditch damage areas, the method further includes: Generate a detection report based on the three-dimensional point cloud data inside the tunnel, the original image, and the marked defect area information, wherein the defect area information includes the defect type, location coordinates, area size, and severity of the defect; The detection report is sent to maintenance personnel or a management system for maintenance decision making and disease treatment.

[0016] In an optional manner, the identifying the drainage ditch area in the grayscale image according to the template matching method further includes: Setting drainage ditch templates with different sizes and resolutions, and searching in the grayscale image to obtain matching results according to a multi-scale template matching method; Remove suspected areas of small area or irregular shape from the search matching results, and merge adjacent areas with high similarity; The matching result is corrected according to the shape, size and grayscale distribution of the suspected area.

[0017] According to the solution provided by the present invention, three-dimensional point cloud data inside the tunnel is collected by a mobile vehicle equipped with a three-dimensional laser scanner, wherein the three-dimensional point cloud data includes tunnel cracks, water leakage, lining surface peeling and drainage ditch damaged areas; the depth value is calculated according to the Z value of the three-dimensional point cloud data, and is mapped to a grayscale range of 0-255 to obtain a grayscale image; the grayscale value mutation area in the grayscale image is identified by an edge detection algorithm to obtain the mutation crack area on the tunnel lining surface; the grayscale value mutation area is gray-segmented to obtain the water leakage area, and the connected domain analysis is performed on the infiltration water area to obtain the area of ​​the water leakage area; the grayscale value mutation area is identified according to the region growing algorithm to obtain the lining surface peeling area; the drainage ditch area in the grayscale image is identified according to the template matching method; in the original image corresponding to the three-dimensional point cloud data, the position and area of ​​the tunnel cracks, water leakage, lining surface peeling and drainage ditch damaged areas are marked. The present invention collects high-precision three-dimensional point cloud data inside the tunnel through a three-dimensional laser scanner, which can comprehensively capture the tunnel cracks, water leakage, lining surface delamination, drainage ditch damage area and other disease information, thereby improving the accuracy and comprehensiveness of detection. The image processing technologies such as edge detection algorithm, grayscale segmentation, region growing algorithm and template matching are used to realize automatic recognition and analysis of the diseased area in the grayscale image, thereby improving the detection efficiency. Through Gaussian filter smoothing and first-order differential operator to calculate the gradient amplitude, the grayscale value mutation area can be accurately identified, and then the mutation crack area on the tunnel lining surface can be identified. The Otsu algorithm is used for grayscale segmentation to effectively distinguish the water leakage area. Through the depth-first search DFS traversal of the binary image, the interconnected white pixels are accurately marked as the same connected domain, thereby accurately calculating the area of ​​the water leakage area. In addition, the drainage ditch area can be accurately identified by the template matching method. The multi-scale template matching method is used to identify the drainage ditch area, which can more accurately match the drainage ditch areas of different sizes and resolutions, thereby improving the accuracy of recognition.

[0018] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings: Figure 1 A schematic diagram showing a flow chart of a tunnel surface defect detection method according to an embodiment of the present invention; Figure 2 A schematic diagram of a manual detection tunnel according to an embodiment of the present invention is shown; Figure 3 A schematic diagram of a tunnel 3D point cloud data according to an embodiment of the present invention is shown. Figure 1 ; Figure 4 A schematic diagram of a tunnel 3D point cloud data according to an embodiment of the present invention is shown. Figure 2 ; Figure 5 A schematic diagram of a tunnel 3D point cloud data according to an embodiment of the present invention is shown. Figure 3 ; Figure 6 The grayscale image and depth image of the same position in the tunnel according to the embodiment of the present invention are shown. Figure 1 ; Figure 7 The grayscale image and depth image of the same position in the tunnel according to the embodiment of the present invention are shown. Figure 2 ; Figure 8 A schematic diagram of identifying deformation defects based on a depth map according to an embodiment of the present invention is shown; Fig. 9 The grayscale image, depth image and high-definition image are accurately matched according to the embodiment of the present invention. Figure 1 ; Fig.10 The grayscale image, depth image and high-definition image are accurately matched according to the embodiment of the present invention. Figure 2 ; Fig.11 The grayscale image disease display intent of an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0020] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0021] Figure 1 FIG. 1 is a flow chart showing a method for detecting tunnel surface defects according to an embodiment of the present invention. Specifically, Figure 1 As shown, the following steps are included: Step S101, collecting three-dimensional point cloud data inside the tunnel by a mobile vehicle equipped with a three-dimensional laser scanner, wherein the three-dimensional point cloud data includes tunnel cracks, water leakage, lining surface peeling, and drainage ditch damage areas.

[0022] In this embodiment, the 3D laser scanner can obtain high-precision spatial point coordinates, including detailed information such as cracks, water leakage, lining surface delamination, and damaged drainage ditch areas in the tunnel. This method does not require direct contact with the tunnel surface, avoiding damage to the tunnel structure caused by detection. A mobile vehicle equipped with a 3D laser scanner can quickly complete the 3D data collection inside the tunnel without interrupting traffic, thereby improving detection efficiency. The 3D point cloud data can fully cover all areas inside the tunnel, including corners and hidden areas that are difficult to reach manually, ensuring the comprehensiveness and accuracy of the detection.

[0023] Specifically, a 3D laser scanner is installed on a moving vehicle, the moving vehicle is started and the 3D laser scanner is turned on. The vehicle moves along the inside of the tunnel, while the scanner continuously emits lasers and receives reflected signals to generate Figures 3 to 5 The 3D point cloud data inside the tunnel is shown. The collected 3D point cloud data is preprocessed by denoising, filtering and other operations.

[0024] Step S102, calculating a depth value according to the Z value of the three-dimensional point cloud data, and mapping it to a grayscale range of 0-255 to obtain a grayscale image.

[0025] In this embodiment, Figures 6 to 10As shown in the figure, the grayscale image retains the depth information of the three-dimensional point cloud data, and the difference in grayscale values ​​can reflect the depth changes in different areas inside the tunnel. First, the X, Y, and Z coordinates of each point are extracted from the three-dimensional point cloud data. Then, a reference plane (such as the bottom of the tunnel or a plane at a specific height) is determined, and the vertical distance from each point to the reference plane, that is, the depth value, is calculated. The calculated depth value is normalized so that it is distributed in the range of 0-1. Then, according to the normalized depth value, it is mapped to the grayscale range of 0-255. According to the mapped grayscale value, the corresponding grayscale image is generated.

[0026] Step S103, identifying the grayscale value mutation area in the grayscale image through the edge detection algorithm to obtain the mutation crack area on the tunnel lining surface; performing grayscale segmentation on the grayscale value mutation area to obtain the leakage area, and performing connected domain analysis on the seepage water area to obtain the area of ​​the leakage area; identifying the grayscale value mutation area according to the region growing algorithm to obtain the lining surface delamination area; and segmenting the drainage ditch area in the grayscale image according to the template matching method.

[0027] In an optional manner, the step of identifying the grayscale value mutation region in the grayscale image by using an edge detection algorithm to obtain the mutation crack region on the tunnel lining surface further comprises: Smoothing the grayscale image using a Gaussian filter, and calculating the gradient amplitude of each pixel in the grayscale image using a first-order differential operator, wherein the gradient amplitude is the rate of change of the grayscale value at the pixel; Determine the pixel points whose gradient amplitude exceeds the preset high threshold as strong edge points, determine the pixel points between the preset low threshold and the preset high threshold as weak edge points, and determine the pixel points below the preset low threshold as non-edge points; The weak edge points are connected to the strong edge points to form an edge profile, thereby obtaining a sudden crack area on the tunnel lining surface.

[0028] In this embodiment, the Gaussian filter can effectively smooth the noise in the image and reduce the interference of noise on the edge detection results. The first-order differential operator can accurately calculate the gradient amplitude of each pixel, that is, the rate of change of the gray value, so as to accurately reflect the edge information in the image. By setting a preset low threshold and a preset high threshold, strong edge points, weak edge points and non-edge points can be flexibly determined to avoid edge omission or false detection that may be caused by a single threshold determination. Weak edge points are connected to strong edge points to form edge contours, which can clearly depict the sudden crack area on the tunnel lining surface.

[0029] For example, a Gaussian filter with a standard deviation of 1.0 is used to smooth the grayscale image to obtain a smoothed image. The Sobel operator is used to calculate the gradient amplitude of each pixel in the smoothed image to obtain a gradient amplitude image, in which the area with higher brightness indicates that the gradient amplitude is larger, that is, the edge information is more obvious. The preset low threshold is set to 50 and the preset high threshold is set to 150. Pixels with a gradient amplitude exceeding 150 are determined as strong edge points. Pixels with a gradient amplitude between 50 and 150 are determined as weak edge points. Pixels with a gradient amplitude below 50 are determined as non-edge points. The edge tracking algorithm is used to connect the weak edge points to the strong edge points to form an edge contour. According to the formed edge contour, the sudden crack area on the tunnel lining surface can be obtained.

[0030] In an optional manner, the calculation formula for the gradient amplitude is:

[0031] in, is the Sobel operator in the horizontal direction, ; is the Sobel operator in the vertical direction, ; is the pixel horizontal coordinate; is the pixel vertical coordinate.

[0032] In this embodiment, the Sobel operator is combined with Gaussian smoothing and differential operations to accurately detect edge information in the image while suppressing noise. By calculating the gradient amplitude in the horizontal and vertical directions, the edge features in the image can be fully captured. The Sobel operator has strong robustness to noise and illumination changes in the image and can work stably in images with different illumination conditions and different noise levels. By calculating the gradient amplitude in the horizontal and vertical directions respectively, the direction information of the edge can be obtained.

[0033] In an optional manner, performing grayscale segmentation on the grayscale value mutation area to obtain the water leakage area further includes: Initialize the weights and means of the foreground and background in the gray value mutation area according to the Otsu algorithm; The inter-class variance is calculated according to the weight and the mean, and the gray value mutation area is segmented into a water leakage area according to the maximum threshold of the inter-class variance.

[0034] In this embodiment, the Otsu algorithm finds the optimal threshold value by maximizing the inter-class variance during the segmentation process, and divides the grayscale image into two parts, the foreground and the background. For example, the grayscale histogram statistics are performed on the grayscale value mutation area to obtain the number of pixels at each grayscale level, and all possible thresholds (such as 0-255) are traversed to divide the grayscale value mutation area into two parts, the foreground and the background. For each threshold, the weight and mean of the foreground and the background are calculated, and the inter-class variance is calculated based on these values. Find the threshold value that maximizes the inter-class variance, and use the maximum threshold value to divide the grayscale value mutation area into a water seepage area (such as a part with a grayscale value greater than 128) and a non-water seepage area (a part with a grayscale value less than or equal to 128).

[0035] In an optional manner, the calculation formula for the between-class variance is:

[0036] in, , Grayscale The ratio of the number of pixels in the image to the total number of pixels in the image, is the preset threshold; ; ; , is the number of gray levels; , is the mean value of the entire image.

[0037] In this embodiment, the quality of segmentation is evaluated by calculating the difference between the foreground and background and the mean of the entire image, which can more accurately reflect the quality of the segmentation result, thereby obtaining a more accurate water leakage area. For example, the number of gray levels L is set to 256 (for an 8-bit grayscale image). All possible thresholds are traversed, starting from T=0, and gradually increasing the value of T until T=255. For each T value, calculate according to the formula , , , , between-class variance . Find The maximum T value. Use the maximum T value to segment the grayscale image into leaking water areas (such as parts with grayscale values ​​greater than or equal to 100) and non-leaking water areas (such as parts with grayscale values ​​less than 100).

[0038] In an optional manner, the performing a connected domain analysis on the infiltrated water area to obtain the area of ​​the leaked water area further comprises: Traversing the binary image in the infiltration water area by depth-first search (DFS), marking interconnected white pixels as the same connected domain; The number of pixels contained in each marked connected domain is calculated to obtain the area of ​​the water leakage region.

[0039] In this embodiment, the image is converted into a binary image, in which the value of white pixels is 1 and the value of black pixels is 0. A tag array of the same size as the image is initialized, and the values ​​of all elements are set to 0, indicating that all pixels have not been visited. The traversal starts from the upper left corner of the image, and when a pixel with a value of 1 and has not been visited is encountered, the DFS traversal is started. During the DFS traversal process, the visited pixels are marked as 1 in the tag array to avoid repeated searches. Whenever a new connected domain is found, a unique label (such as 1, 2, 3, etc.) is assigned to the connected domain, and the number of pixels contained in the connected domain is recorded. After the traversal is completed, the number of pixels contained in each connected domain is counted according to the tag array and the connected domain label. Assuming that there are three connected domains in the image, containing 4, 2 and 1 white pixels respectively, the areas of these three connected domains are 4, 2 and 1 respectively, and these three areas are used as the areas of the water leakage area.

[0040] In an optional manner, the step of identifying the gray value mutation region according to the region growing algorithm to obtain the lining surface delamination region further comprises: Starting from the initial seed point of the gray value mutation area, searching for its neighboring pixels; Determine whether the neighborhood pixel meets the adding condition according to the growth criterion, if so, add it to the current area and regard it as a new seed point; During the region growing process, the gray value difference of each pixel when it is added to the region is recorded. When the gray value difference of each pixel exceeds a preset threshold, the pixel is located in the gray value mutation region; The lining surface delamination area is determined according to the gray value mutation area.

[0041] In this embodiment, an initial seed point located near the layering area is selected in the image. Starting from the initial seed point, its 8 neighboring pixels are searched. According to the criterion that the gray value difference is less than 5, the neighboring pixels that meet the conditions are added to the current area and regarded as new seed points. Repeat this process until there are no neighboring pixels that meet the conditions to be added to the current area. During the region growing process, the gray value difference of each pixel when it is added to the region is recorded. For example, the gray value difference of a certain pixel when it is added to the region is 7. Set the preset threshold to 6. When the gray value difference of a certain pixel exceeds 6, it is considered that the pixel is located in the gray value mutation area. Pixels with a gray value difference of 7 are identified as being located in the gray value mutation area. The layering area of ​​the lining surface is determined according to the position and shape of the gray value mutation area. The boundary of the layering area can be optimized by morphological processing (such as expansion) to obtain the final layering area recognition result.

[0042] In an optional manner, the separating the drainage ditch area in the grayscale image according to the template matching method further includes: Presetting a drainage ditch area as a template, wherein the template includes the shape, size and grayscale distribution of the drainage ditch; An area similar to the template is searched in the grayscale image, and an area whose similarity exceeds a preset threshold is determined as a drainage ditch area.

[0043] In this embodiment, a template image is made according to the shape, size and grayscale distribution of the drainage ditch. For example, the template image is a long strip area, and its grayscale value distribution is similar to that of the drainage ditch. The input grayscale image is denoised and contrast enhanced to ensure that the resolution of the grayscale image is consistent with the resolution of the template image. The template is slid in the grayscale image to calculate the normalized mutual information between the template and each position in the image as a similarity measure. The preset threshold is set to 0.8 (indicating that the match is successful when the similarity reaches more than 80%), and the area with a similarity exceeding 0.8 is determined as a drainage ditch area.

[0044] In an optional manner, the identifying the drainage ditch area in the grayscale image according to the template matching method further includes: Setting drainage ditch templates with different sizes and resolutions, and searching in the grayscale image to obtain matching results according to a multi-scale template matching method; Remove suspected areas of small area or irregular shape from the search matching results, and merge adjacent areas with high similarity; The matching result is corrected according to the shape, size and grayscale distribution of the suspected area.

Claims

1. A method for detecting apparent defects in a tunnel, characterized in that: include: Collecting 3D point cloud data of the interior of the tunnel by a mobile vehicle equipped with a 3D laser scanner, wherein the 3D point cloud data includes tunnel cracks, water leakage, lining surface delamination, and drainage ditch damage areas; Calculate the depth value according to the Z value of the three-dimensional point cloud data, and map it to a grayscale range of 0-255 to obtain a grayscale image; The grayscale mutation region in the grayscale image is identified by an edge detection algorithm to obtain the mutation crack region on the tunnel lining surface; the grayscale mutation region is segmented to obtain the leakage region, and the connected domain analysis is performed on the seepage region to obtain the area of ​​the leakage region; the grayscale mutation region is identified by a region growing algorithm to obtain the lining surface layer region; the drainage ditch region in the grayscale image is segmented by a template matching method; In the original image corresponding to the three-dimensional point cloud data, the positions and areas of the tunnel cracks, water leakage, lining surface peeling and drainage ditch damage areas are marked.

2. The method according to claim 1, characterized in that The step of identifying the grayscale value mutation region in the grayscale image by using an edge detection algorithm to obtain the mutation crack region on the tunnel lining surface further comprises: Smoothing the grayscale image using a Gaussian filter, and calculating the gradient amplitude of each pixel in the grayscale image using a first-order differential operator, wherein the gradient amplitude is the rate of change of the grayscale value at the pixel; Determine the pixel points whose gradient amplitude exceeds the preset high threshold as strong edge points, determine the pixel points between the preset low threshold and the preset high threshold as weak edge points, and determine the pixel points below the preset low threshold as non-edge points; The weak edge points are connected to the strong edge points to form an edge profile, thereby obtaining a sudden crack area on the tunnel lining surface.

3. The method according to claim 2, characterized in that The calculation formula of the gradient amplitude is: in, is the Sobel operator in the horizontal direction, ; is the Sobel operator in the vertical direction, ; is the pixel horizontal coordinate; is the pixel vertical coordinate.

4. The method according to claim 1, characterized in that The grayscale segmentation of the grayscale value mutation area to obtain the water leakage area further includes: Initialize the weights and means of the foreground and background in the gray value mutation area according to the Otsu algorithm; The inter-class variance is calculated according to the weight and the mean, and the gray value mutation area is segmented into a water leakage area according to the maximum threshold of the inter-class variance.

5. The method according to claim 4, characterized in that The calculation formula of the inter-class variance is: in, , Grayscale The ratio of the number of pixels in the image to the total number of pixels in the image, is the preset threshold; ; ; , is the number of gray levels; , is the mean value of the entire image.

6. The method according to claim 1, characterized in that The connecting domain analysis of the water seepage area to obtain the area of ​​the water leakage area further comprises: Traversing the binary image in the infiltration water area by depth-first search (DFS), marking interconnected white pixels as the same connected domain; The number of pixels contained in each marked connected domain is calculated to obtain the area of ​​the water leakage region.

7. The method according to claim 1, characterized in that The step of identifying the gray value mutation region according to the region growing algorithm to obtain the lining surface delamination region further comprises: Starting from the initial seed point of the gray value mutation area, searching for its neighboring pixels; Determine whether the neighboring pixel meets the adding condition according to the growth criterion, if so, add it to the current area and regard it as a new seed point; During the region growing process, the gray value difference of each pixel point when it is added to the region is recorded. When the gray value difference of each pixel point exceeds a preset threshold, the pixel point is located in the gray value mutation region; The lining surface delamination area is determined according to the gray value mutation area.

8. The method according to claim 1, characterized in that The drain ditch area in the grayscale image according to the template matching method further includes: Presetting a drainage ditch area as a template, wherein the template includes the shape, size and grayscale distribution of the drainage ditch; An area similar to the template is searched in the grayscale image, and an area whose similarity exceeds a preset threshold is determined as a drainage ditch area.

9. The method according to claim 1, characterized in that: After marking the locations and areas of the tunnel cracks, water leakage, lining surface delamination, and drainage ditch damage areas, the method further includes: Generate a detection report based on the three-dimensional point cloud data inside the tunnel, the original image, and the marked defect area information, wherein the defect area information includes the defect type, location coordinates, area size, and severity of the defect; The detection report is sent to maintenance personnel or a management system for maintenance decision making and disease treatment.

10. The method according to claim 8, characterized in that The identifying of the drainage ditch area in the grayscale image according to the template matching method further comprises: Setting drainage ditch templates with different sizes and resolutions, and searching in the grayscale image to obtain matching results according to a multi-scale template matching method; Remove suspected areas of small area or irregular shape from the search matching results, and merge adjacent areas with high similarity; The matching result is corrected according to the shape, size and grayscale distribution of the suspected area.

Images