Tunnel lining development view-based method and apparatus for identifying spatial distribution characteristics of cracks

Abstract

A tunnel lining development view-based method and apparatus for identifying spatial distribution characteristics of cracks, relating to the technical field of tunnel engineering. On the basis of a tunnel lining crack section analysis-intended area panoramic development view, in combination with a crack defect deep learning model and a conventional image processing method, the present invention can effectively reduce data redundancy, and efficiently and accurately implement multi-scale crack identification. Deep learning implements crack defect positioning and extraction at a large scale, conventional image processing implements crack defect parameter extraction at a small scale, and then a kilometer-scale tunnel crack spatial distribution and parameter display model is constructed. The present invention can intuitively obtain overall spatial distribution characteristics and related quantitative parameters of tunnel lining crack defects in a three-dimensional space at a kilometer level, such as a crack length, a crack width, a mileage position where a crack is present, and a cross-sectional position where a crack is present, thereby providing an important basis for subsequent specific analysis of the crack defects.

Description

Tunnel lining unfolding map crack spatial distribution feature identification method and device TECHNICAL FIELD

[0001] The present application relates to the technical field of tunnel engineering, and in particular to a tunnel lining unfolding map crack spatial distribution feature identification method and device. BACKGROUND

[0002] With the continuous advancement of highway, railway and urban rail transit construction in China, tunnel engineering in China has entered a new period of equal emphasis on construction and maintenance, and tunnel lining crack damage problems have attracted much attention. As an index that can directly reflect the stability and safety of tunnel lining, the efficient and accurate acquisition of the overall spatial distribution characteristics of crack damage is of great significance to ensure the safe operation of tunnels.

[0003] In engineering, the detection of tunnel lining crack damage is currently generally carried out by manual drawing, which has a large error and may mislead further research and analysis. In recent years, thanks to the development of computer and artificial intelligence technology, many algorithms and models have been able to realize the rapid and intelligent detection of tunnel lining crack damage. For example, based on deep learning algorithm, cracks in tunnel lining images are identified; based on traditional image processing, the width, length and number of cracks are obtained; based on image stitching technology, the position information of tunnel lining crack damage is obtained. However, these detection methods cannot efficiently, accurately, intuitively and comprehensively obtain the overall spatial distribution characteristics of tunnel lining crack damage.

[0004] In addition, the current research lacks in the specific information display of tunnel damage, especially in the intuitive acquisition of the overall spatial distribution characteristics of tunnel lining crack damage in three-dimensional space at the kilometer level and related quantitative parameters such as crack length, crack width, crack location and crack cross-section location (such as vault / roof, haunch / side wall and arch foot), which is not conducive to the subsequent analysis and judgment of tunnel conditions.

[0005] SUMMARY

[0006] In order to overcome the deficiencies in the background art, the present application discloses a tunnel lining unfolding map crack spatial distribution feature identification method and device.

[0007] To achieve the above-mentioned application purposes, the present application adopts the following technical solutions:

[0008] In the first aspect, the present application provides a tunnel lining unfolding map crack spatial distribution feature identification method and device, comprising the following steps:

[0009] S1, obtaining a three-dimensional grid model of the tunnel lining crack section to be analyzed;

[0010] S2. Perform a panoramic unfolding of the three-dimensional mesh model of the proposed analysis area of ​​the tunnel lining cracked section and the tunnel lining design outline to obtain a panoramic unfolding map of the proposed analysis area of ​​the tunnel lining cracked section.

[0011] S3. Using a pre-trained crack damage identification model, crack damage is identified in the first bottom block of the panoramic unfolded map of the cracked section of the tunnel lining, resulting in the second bottom block. Based on the distribution of crack areas in the second bottom block, misidentified crack areas are removed. Then, the second bottom block is stitched together step by step to obtain the overall distribution characteristics of tunnel crack damage.

[0012] S4. After binarizing the image of the second bottom block, the spatial distribution characteristics of cracks in the tunnel fracture section are obtained by dilatation erosion and skeleton extraction.

[0013] S5. Based on the spatial distribution characteristics of cracks in the tunnel's damaged sections and the results of crack regions, construct a kilometer-level tunnel crack spatial distribution and parameter display model.

[0014] Specifically, in step S3, the crack damage identification model is used to identify crack damage in the first bottom block of the panoramic unfolded image of the proposed analysis area of ​​the tunnel lining cracked section, and the second bottom block is obtained. The storage method of the panoramic unfolded image of the proposed analysis area of ​​the tunnel lining cracked section is hierarchical storage, including [0,1,...,N] levels. Each level image consists of 4 images of the lower level. The Nth level image is obtained as the first bottom block. The crack damage identification model is used to identify crack damage in the first bottom block to obtain the second bottom block containing the identified crack area results.

[0015] Specifically, step S3, which involves removing misidentified crack areas based on the distribution of crack area results in the second bottom layer block, includes:

[0016] Obtain the pixel coordinates of the center point of each crack region in the second bottom layer block, and the number of rows and columns of pixels in the crack region, respectively, and represent them as multiple one-dimensional arrays [l c ,h c The crack region list L is stored in the form of [,dl,dh]. lab In, among which: l c h represents the column coordinates of the pixel point at the center of the crack region. c dl represents the row coordinates of the pixel center point of the crack region, dh represents the number of columns of the crack region, and dh represents the number of rows of the crack region.

[0017] List L for each crack region lab Iterate through the list and read each L in turn. labAll crack regions contained in the list; traverse all crack regions, magnify the length and width of the current crack region by a preset determination multiple r as a determination region, then find other crack regions in the determination region, if there are other crack regions in the determination region, retain the crack region, if not, eliminate the crack region.

[0018] Specifically, if there are other crack regions in the determination region, the crack region is retained, and if not, the crack region is eliminated. Specifically, traverse all crack regions except the current crack region, compare the center point of each remaining crack region with the boundary of the determination region in turn, if the center point is within the boundary of the determination region, it is considered that there are other crack regions in the determination region, and the region is retained, otherwise, the current crack region is an isolated region and is eliminated.

[0019] Specifically, step S3 further comprises: after obtaining the second bottom layer block, performing gap type identification on the second bottom layer block; based on the crack region center points in the identification result, traversing the image of the second bottom layer block in the pixel range of each second bottom layer block using a square traversal frame of a*a pixels; in the traversal process, if the number of crack region center points in the traversal frame range is greater than a preset crack number threshold f, it is considered that the image is a net-shaped crack; otherwise, it is considered that the image is a strip-shaped crack.

[0020] Specifically, step S4 specifically comprises the following steps:

[0021] S41, using a slicing algorithm in a numpy array to extract a crack region image according to the boundary of the crack region in the second bottom layer block;

[0022] S42, obtaining a crack region binary image of the crack region image through an image binaryzation algorithm;

[0023] S43, splicing the crack region binary images belonging to the same second bottom layer block according to coordinates to obtain a binary image bottom layer block;

[0024] S44, connecting the interruption points of adjacent cracks in the binary image bottom layer block through a dilation and erosion algorithm, and then extracting a crack region skeleton through a skeleton extraction algorithm to obtain a crack region skeleton block;

[0025] The parameter selection and running times of the dilation and erosion algorithm and the skeleton extraction algorithm are different according to the gap type of the second bottom layer block; wherein the net-shaped crack adopts small structure element N times dilation and erosion and skeleton extraction operation, and the strip-shaped crack adopts large structure element single-time dilation and erosion and skeleton extraction operation; the structure element is only expanded by a single layer of pixels as a small structure element, and the structure element is expanded by two layers of pixels or more as a large structure element.

[0026] S45, the crack area skeleton block is spliced step by step to obtain the crack spatial distribution characteristics of the tunnel crack section.

[0027] Specifically, the step S5 specifically comprises:

[0028] S51, the conversion coefficient a = d / d of the actual distance and the pixel is calculated. p , unit: m / px; d is the length of the design contour inner surface cross section image acquisition range, unit: m; d p is the panorama development map width, unit: px;

[0029] S52, the horizontal and vertical cross section positions where the crack is located are determined according to the crack skeleton endpoint coordinates in the crack area skeleton block; specifically,

[0030] When the crack skeleton endpoint vertical coordinate y satisfies y jy1 or y jy2 , it represents that the crack is located at the springing position,

[0031] When the crack skeleton endpoint vertical coordinate y satisfies y jy1 <y<y yd1 or y yd2 <y<y jy2 , it represents that the crack is located at the haunch position,

[0032] When the crack skeleton endpoint vertical coordinate y satisfies y yd1 <y<y yd2 , it represents that the crack is located at the crown position.

[0033] Wherein y jy1 and y jy2 are the springing and haunch demarcation points, y yd1 and y yd2 are the haunch and crown demarcation points, and the crack skeleton center point horizontal coordinate x represents the vertical cross section position where the crack is located.

[0034] S53, the overall crack length in the image is calculated according to the crack area skeleton block, and a kilometer-level tunnel crack spatial distribution and crack length display model is constructed.

[0035] S54, the crack average width is calculated according to the crack area in the second bottom layer block, and a kilometer-level tunnel crack spatial distribution and crack average width display model is constructed.

[0036] Specifically, the step S53 specifically comprises: multiplying the number of black pixel points n' ij of the crack area skeleton block p' ij by the conversion coefficient a, and calculating the crack length l' ij of each block.Then, by splicing adjacent cracks, the length of the entire crack is obtained, and finally a kilometer-level tunnel crack spatial distribution and crack length display model is constructed.

[0037] Specifically, step S54 is as follows:

[0038] First, calculate the second bottom-level block a′ ij The minimum length d′ of the diagonal of the bounding rectangle of the black pixel ij-c Calculate the crack region a′ ij-c Number of black pixels n′ ij-c The ratio method was used to calculate the crack region a′. ij-c The average crack width w′ in ij-c As shown in the following formula:

[0039] w′ ij-c =a·n′ ij-c / d′ ij-c Where c is the second bottom layer block a′ ij The identified crack areas are numbered;

[0040] Then calculate block a′ according to the following formula. ij Average crack width within Finally, a model was constructed to display the spatial distribution and average width of cracks in a kilometer-level tunnel.

[0041] Secondly, the present invention provides a device for identifying the spatial distribution characteristics of cracks in tunnel lining based on a developed diagram, comprising the following units:

[0042] Three-dimensional reconstruction unit is used to obtain a three-dimensional mesh model of the area to be analyzed in the tunnel lining cracked section;

[0043] The panoramic unfolding construction unit is used to unfold the three-dimensional mesh model of the proposed analysis area of ​​the tunnel lining cracked section and the tunnel lining design outline in a panoramic manner to obtain the panoramic unfolding map of the proposed analysis area of ​​the tunnel lining cracked section.

[0044] The overall distribution feature construction unit of crack damage uses a pre-trained crack damage identification model to identify crack damage in the first bottom block of the panoramic unfolded map of the cracked section of the tunnel lining, obtains the second bottom block, and removes misidentified crack areas based on the distribution of crack areas in the second bottom block. Then, the second bottom block is stitched together step by step to obtain the overall distribution feature of tunnel crack damage.

[0045] The crack spatial distribution feature construction unit of the cracked section is used to binarize the image of the second bottom block and then obtain the crack spatial distribution feature of the tunnel cracked section through dilatation erosion and skeleton extraction.

[0046] The crack area result display unit is used to construct a kilometer-level tunnel crack spatial distribution and parameter display model based on the spatial distribution characteristics of cracks in the tunnel damage section and the crack area results.

[0047] The present invention provides a method for identifying the spatial distribution characteristics of cracks in tunnel lining based on the unfolded diagram, comprising the following steps: S1, performing three-dimensional reconstruction on the image of the proposed analysis area of ​​the cracked section of the tunnel lining to obtain a three-dimensional mesh model of the proposed analysis area of ​​the cracked section of the tunnel lining; S2, performing panoramic unfolding of the three-dimensional mesh model of the proposed analysis area of ​​the cracked section of the tunnel lining and the tunnel lining design outline to obtain a panoramic unfolded diagram of the proposed analysis area of ​​the cracked section of the tunnel lining; S3, training a crack disease identification model on the image of the proposed analysis area of ​​the cracked section of the tunnel lining to obtain a crack disease identification model; S4, using the crack disease identification model to identify the cracked section of the tunnel lining. The first bottom block of the panoramic unfolded image of the tunnel lining cracked section is used to identify crack defects, and a second bottom block is obtained. Based on the distribution of crack areas in the second bottom block, misidentified crack areas are removed. Then, the second bottom block is stitched together step by step to obtain the overall distribution characteristics of tunnel crack defects. S5. After image binarization of the second bottom block, the spatial distribution characteristics of cracks in the tunnel cracked section are obtained through expansion erosion and skeleton extraction. S5. Based on the spatial distribution characteristics of cracks in the tunnel cracked section and the crack area results, a kilometer-level tunnel crack spatial distribution and parameter display model is constructed.

[0048] This invention, based on a panoramic unfolded map of the proposed analysis area of ​​the tunnel lining cracked section, and combining a deep learning model of crack damage with traditional image processing methods, can effectively reduce data redundancy and efficiently and accurately identify cracks at multiple scales. Deep learning enables the localization and extraction of crack damage at a large scale, while traditional image processing extracts crack damage parameters at a small scale. This invention can intuitively obtain the overall spatial distribution characteristics and related quantitative parameters of tunnel lining crack damage in three-dimensional space at the kilometer level, such as crack length, crack width, crack location at kilometer level, and crack location in cross-sections (e.g., arch crown / top slab, arch waist / sidewall, arch foot), providing an important foundation for subsequent specific analysis of crack damage and is of great significance.

[0049] Furthermore, the method for eliminating misidentified areas in this invention further improves the accuracy of identifying tunnel cracked areas, avoids unnecessary on-site inspections, and reduces maintenance costs.

[0050] Furthermore, this invention uses a hierarchical storage method to store image data of different details at different levels of the panoramic unfolded image. When processing and analyzing the subsequent kilometer-level tunnel images, the higher-level images can be preprocessed first to identify areas of interest or to perform preliminary analysis. Then, these areas can be gradually zoomed in to view more details and perform detailed processing. This can improve processing efficiency and reduce unnecessary consumption of computing resources.

[0051] In addition, the extraction process of the tunnel crack section crack space distribution characteristics of the present application uses binarization first, so that the image data quantity is smaller, the processing speed is faster, and large-scale image analysis and processing can be effectively carried out. Through morphological operations such as expansion and corrosion, noise and small interference elements can be removed, and broken cracks can be connected, so that the crack characteristics are more obvious. Through skeleton extraction, the basic shape and structure of the crack can be obtained, which is more convenient for analyzing the spatial distribution characteristics such as the direction, width and length of the crack, and the efficiency of crack analysis of the crack section is effectively improved.

[0052] In addition, the present application can ensure the original morphology of the network crack and the connectivity of the strip crack as much as possible through different parameter selection and operation times of the expansion and corrosion algorithm and the skeleton extraction algorithm according to different crack categories. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0054] Fig. 1 is a flowchart of a crack space distribution feature recognition method based on a tunnel lining development drawing according to an embodiment of the present application;

[0055] Fig. 2 is a marker plate style diagram when control points are laid out according to an embodiment of the present application, wherein Fig. 2(a) is a random style 1 of the calibration plate, and Fig. 2(b) is a random style 2 of the calibration plate;

[0056] Fig. 3 is a rear intersection method diagram after total station measurement according to an embodiment of the present application;

[0057] Fig. 4 is a three-dimensional grid model diagram of a tunnel lining crack section analysis area according to an embodiment of the present application;

[0058] Fig. 5 is a three-dimensional grid model diagram of a tunnel lining design contour according to an embodiment of the present application;

[0059] Fig. 6 is a panoramic development drawing of a tunnel lining crack section analysis area according to an embodiment of the present application;

[0060] Fig. 7 is an image labeling diagram of a tunnel lining crack section analysis area according to an embodiment of the present application;

[0061] Fig. 8 is a panoramic development drawing storage diagram according to an embodiment of the present application; wherein Fig. 8(a) is a 17-level picture example, and Fig. 8(b) is a 16-level picture example;

[0062] Fig. 9 is a schematic diagram of a tunnel crack site panoramic development block crack identification provided by an embodiment of the present application, wherein Fig. 9(a) is a kind of crack region identification situation, and Fig. 9(b) is another crack region identification situation;

[0063] Fig. 10 is a schematic diagram of determining crack category by traversing the number of crack region center points in a block provided by an embodiment of the present application, wherein Fig. 10(a) is a strip crack instance, and Fig. 10(b) is a mesh crack instance;

[0064] Fig. 11 is a schematic diagram of the principle of screening isolated crack regions provided by an embodiment of the present application;

[0065] Fig. 12 is a schematic diagram of the result of eliminating isolated regions provided by an embodiment of the present application, wherein Fig. 12(a) is before the elimination of isolated regions, and Fig. 12(b) is after the elimination of isolated regions;

[0066] Fig. 13 is a schematic diagram of the overall distribution characteristics of tunnel crack damage provided by an embodiment of the present application;

[0067] Fig. 14 is a process diagram for extracting the spatial distribution characteristics of tunnel crack sections provided by an embodiment of the present application, wherein Fig. 14(a) is a block crack identification result, Fig. 14(b) is a crack damage region identified in the block, Fig. 14(c) is a binary image of the crack damage region identified in the block, and Fig. 14(d) is a binary image of the crack damage in the block;

[0068] Fig. 15 is a flowchart for extracting the skeleton image of crack damage in a block in a tunnel crack site panoramic development diagram provided by an embodiment of the present application, wherein Fig. 15(a) is a binary image of crack damage in the block, Fig. 15(b) is an image processed by an inflation erosion algorithm, and Fig. 15(c) is an image processed by a skeleton extraction algorithm;

[0069] Fig. 16 is a skeleton crack identification result of a panoramic development diagram provided by an embodiment of the present application;

[0070] Fig. 17 is a model for displaying the spatial distribution and parameters of a kilometer-level tunnel crack provided by an embodiment of the present application, wherein Fig. 17(a) is a model for displaying the spatial distribution and crack length of a kilometer-level tunnel crack, and Fig. 17(b) is a model for displaying the spatial distribution and average width of a kilometer-level tunnel crack;

[0071] Fig. 18 is a schematic diagram of a device for identifying the spatial distribution characteristics of a tunnel lining development diagram crack damage according to an embodiment of the present application;

[0072] Fig. 19 is a schematic diagram of a device for identifying the spatial distribution characteristics of a tunnel lining development diagram crack damage according to an embodiment of the present application. DETAILED DESCRIPTION

[0073] The application can be explained in detail by the following examples, the purpose of the disclosure is to protect all technical improvements within the scope of the application, and in the description of the application, it should be understood that if the terms "upper", "lower", "front", "back", "left", "right" and the like indicate the orientation or position relationship, only correspond to the drawings of the present application, in order to facilitate the description of the application, and not indicate or imply that the device or element must have a particular orientation.

[0074] Example one

[0075] Referring to FIG. 1, the embodiment provides a tunnel lining unfolding diagram crack damage space distribution feature identification method, including the following steps:

[0076] S1, obtaining a three-dimensional grid model of a tunnel lining crack section analysis area;

[0077] Specifically, the step S1 specifically includes:

[0078] S11, image acquisition in the tunnel lining crack section analysis area;

[0079] The embodiment obtains a tunnel lining image by laying control points in the tunnel lining crack section analysis area;

[0080] When laying control points in the tunnel lining image acquisition area, first, 4-6 marker plates are arranged in the area to be photographed, as shown in FIG. 2, and the marker plates should be evenly distributed in the tunnel lining image acquisition area. When arranging the marker plates, use nails (or adhesive tape) to fix the marker plates on the wall, and need to ensure the stability of the marker plates, and they will not move during photographing and measuring. Then, the coordinates of the calibration plate are measured by using a total station, and the center coordinates of the calibration plate are obtained as control points for converting the construction coordinate system;

[0081] When taking pictures, try to ensure that the pictures are perpendicular to the tunnel surface, and ensure that the overlapping degree of any two consecutive pictures is greater than 50%, and the pictures containing the same control point in any direction are not less than three.

[0082] In the embodiment, the total station is erected by exemplarily selecting the positions of each control point and the known coordinates of the rear view point. The direction and distance of the known coordinates of the rear view point 1 and the rear view point 2 are measured by using the total station, and the total station is set by the resection method, as shown in FIG. 3, and then the control point coordinates are measured in sequence by using the total station, and recorded according to the marker plate number.

[0083] S12, obtaining a three-dimensional grid model of a tunnel lining crack section analysis area based on a three-dimensional reconstruction method according to the obtained tunnel lining crack section analysis area image.

[0084] The embodiment is based on a motion recovery structure algorithm SFM to reconstruct a three-dimensional sparse point cloud {V M} of the image, based on a PMVS multi-view matching algorithm to reconstruct a three-dimensional dense point cloud {V′ P} of the image, and based on control point coordinates to obtain a three-dimensional dense point cloud {V P} in a construction coordinate system through rotation, translation and scaling:

[0085] {V P} = {V′ P} · [R | t] · s;

[0086] wherein R is a rotation matrix, t is a translation vector, and s is a scaling factor; the coordinates of the control point in the reconstruction coordinate system are p′ i , and the corresponding point in the construction coordinate system is p i , and the corresponding transformation parameters are solved as follows:

[0087] wherein n represents the number of control points.

[0088] Based on a Delaunay triangulation algorithm, a three-dimensional mesh model S({T}, {V P}) of the analysis region of the tunnel lining crack section is reconstructed according to the three-dimensional dense point cloud {V P} in the construction coordinate system, and {T} represents the triangulation relationship of the vertex set {V P}, as shown in FIG. 4.

[0089] S2, panoramic unfolding is performed on the three-dimensional mesh model of the analysis region of the tunnel lining crack section and the tunnel lining design contour to obtain a panoramic unfolding diagram of the analysis region of the tunnel lining crack section.

[0090] Specifically, step S2 includes the following steps:

[0091] S21, the streamline S(η) and the trace line L(ξ) of the tunnel are obtained according to the tunnel lining design contour;

[0092] The method for establishing the three-dimensional mesh model of the tunnel lining design contour in the embodiment is as follows: a three-dimensional sweeping method is adopted, the streamline S(η) adopts the design section contour line, and if there is no design file, the region contour line is used instead, and the trace line L(ξ) adopts the tunnel design curve, and if there is no design file, the spring line of the actual contour three-dimensional mesh model or the horizontal line with a landmark on the arch wall is used instead.

[0093] S22, the three-dimensional mesh model of the tunnel lining design contour is established according to the obtained streamline S(η) and trace line L(ξ);

[0094] Firstly, the streamline is equally divided with the trace line, the coordinate of each equally divided point is obtained, the more the equally divided points are divided, the higher the calculation result accuracy is, the equally divided points of the streamline are swept according to the direction of the trace line, and a three-dimensional grid model of the tunnel lining design contour is obtained, and the sweeping result is shown in Fig. 5.

[0095] The plane grid points G of the design section contour line are divided i,j (ξ i ,η j ), the space coordinates M i,j (x ij ,y ij ,z ij ) are obtained by sweeping, which can be calculated by the following formula:

[0096] Wherein, N C represents the number of columns of the divided grid, N R represents the number of rows of the divided grid, the more the number of rows and columns is divided, the higher the calculation result accuracy is;

[0097] S23, the three-dimensional grid model of the tunnel lining crack damage section tentative analysis area obtained in S12 is projected to the three-dimensional grid model of the tunnel lining design contour obtained in S22, and a panoramic development drawing of the tunnel lining crack damage section tentative analysis area is obtained, as shown in Fig. 6.

[0098] S3, using the pre-trained crack damage identification model to identify the crack damage of the first bottom layer block of the tunnel lining crack damage section tentative analysis area panoramic development drawing, obtaining the second bottom layer block, and according to the distribution of the crack area result in the second bottom layer block, the misidentified crack area is removed, and then the second bottom layer block is spliced to obtain the overall distribution characteristics of the tunnel crack damage;

[0099] Specifically, the training process of the pre-trained crack damage identification model includes the following steps:

[0100] The tunnel lining crack damage section tentative analysis area image is used to mark the cracks of the training set and the verification set;

[0101] The cracks of the photographed similar photos are marked, the marking is performed by using the labelimg software, and the marking is in the yolov5 format, as shown in Fig. 7.

[0102] The model is trained according to the marked training set and verification set.

[0103] The yolov5 model is trained according to the marked photos, and the yolov5l model is used in this embodiment.

[0104] As to how to use the marked photos to train the yolov5 model, it is a prior art, and will not be described here.

[0105] Specifically, the step S3 adopts the crack disease identification model to identify the crack disease of the first bottom layer block of the tunnel lining crack section analysis area panoramic development map, and obtains the second bottom layer block. Specifically, the storage mode of the tunnel lining crack section analysis area panoramic development map is hierarchical storage, including [0, 1,..., N] levels, each level image is composed of 4 low-level images, and the N-level image is obtained as the first bottom layer block; the crack disease identification model is used to identify the crack disease of the first bottom layer block, as shown in FIG. 9, and the second bottom layer block containing the identified crack area result is obtained.

[0106] The data type of the crack area result is a txt file, and the data information contained includes a crack area center point, a crack area length and a crack area width.

[0107] First, the first bottom layer block a of the tunnel crack site analysis area panoramic development map is obtained ij ;

[0108] The storage mode of the panoramic development map is hierarchical storage, from the highest level 0 to the lowest level N, each level image is composed of four low-level images, and each level image size is 512*512 pixels, as shown in FIG. 8, wherein the N-level image is the bottom layer image of the panoramic development map, and a single image contains the most information, which is named as the first bottom layer block a ij . Wherein, (i+1) represents the column position of the block, (j+1) represents the row position of the block. Based on this storage mode, the panoramic development map can display the image level by level during zooming, i represents the column index of the block, and j represents the row index of the block.

[0109] The value of N in the embodiment can be selected according to specific requirements, preferably, the value of N in the embodiment is 17.

[0110] For example, assuming that the k-level image number is (m k , n k ), wherein m represents the (m+1) column of the panoramic image, and n represents the (n+1) row of the panoramic image.

[0111] Then the k+1-level image number is calculated according to the following formula:

[0112] For the first bottom layer block a ij , the model trained in S32 is used to identify the crack disease, as shown in FIG. 9, and the second bottom layer block a′ ij is obtained after the crack disease identification.

[0113] Specifically, after the second bottom layer block is obtained, a gap category of the second bottom layer block is identified;

[0114] Firstly, based on the crack region center points in the identification result, a traversal frame of a*a pixels is adopted to traverse the image of the second bottom layer block in the pixel range of each second bottom layer block; in the traversal process, if the number of crack region center points in the traversal frame range is greater than a preset crack number threshold f, it is considered that the image is a net-shaped crack; otherwise, it is considered that the image is a strip-shaped crack, as shown in FIG. 10.

[0115] FIG. 10(a) is a traversal case discrimination case corresponding to FIG. 9(a), indicating a strip-shaped crack, and FIG. 10(b) is a traversal discrimination case corresponding to FIG. 9(b), indicating a net-shaped crack.

[0116] The values of the preset crack number threshold f and the traversal frame width a are selected after testing and summarizing, and preferably, f = 5 is selected in the embodiment, and the value of a is 0.3 times the pixel value of the width of the second bottom layer block image.

[0117] Specifically, the step S3 and the distribution of the crack region result in the second bottom layer block are specifically included in:

[0118] Firstly, the crack identification mark is read from the identified crack region result data, the pixel coordinates of the center points of each crack region in the second bottom layer block are obtained, and the row and column pixel numbers of the crack region are obtained, which are saved in the crack region list L in the form of a plurality of one-dimensional arrays [l c ,h c ,dl,dh]. lab , wherein: l c represents the pixel column coordinates of the center points of the crack region, h c represents the pixel row coordinates of the center points of the crack region, dl represents the number of columns of the crack region, and dh represents the number of rows of the crack region.

[0119] Each crack region list L lab is traversed, and all crack regions contained in each L lab list are read in sequence; the length and width of the current crack region are enlarged by a preset judgment multiple r as a judgment region, and then other crack regions in the judgment region are searched; if there is other crack region in the judgment region, the crack region is retained, and if there is no other crack region, the crack region is removed.

[0120] Specifically, if there is another crack region in the determination region, the crack region is retained, and if there is not, the crack region is removed. Specifically, all crack regions except the current crack region are traversed, and the center point of each remaining crack region is compared with the boundary of the determination region in turn. If the center point is within the boundary of the determination region, it is considered that there is another crack region in the determination region, and the region is retained. Otherwise, it is considered that the current crack region is an isolated region, and the region is removed.

[0121] For each L lab list, all crack regions contained in the L lab list are read in turn. For each crack region i, other crack regions are searched in a determination region with a predetermined determination multiple r of the length and width of the region. First, the boundary of the determination region is constructed.

[0122] Wherein: i represents the crack region number; left i , right i , up i , down i respectively represent the left, right, top and bottom boundaries of the determination region of the i-th crack region.

[0123] The value of the predetermined determination multiple r is an odd number greater than or equal to 3, which can be selected according to actual needs. In the embodiment, the value of r is 3.

[0124] Then, for the i-th crack region, the center point of each remaining crack region j is judged in turn. If at least one j satisfies the following formula:

[0125] It is considered that there is another crack region in the determination region of the i-th crack region, and the i-th crack region is retained. Otherwise, it is considered that the i-th crack region is an isolated region, which is a misidentified region, and the region is removed. The principle is shown in FIG. 11.

[0126] Wherein: j represents the traversal result of the remaining crack region number when the i-th crack region is considered. Through testing, when r is 3, good isolated crack region removal effect is obtained, and the specific removal effect is shown in FIG. 12.

[0127] Through the above method, the misidentified region can be effectively removed, the accuracy of tunnel crack damage region identification is further improved, unnecessary on-site inspection is avoided, and maintenance cost is reduced.

[0128] The second bottom layer block a′ ij after the misidentified region is removed is spliced level by level, and finally the overall distribution characteristics of the tunnel crack damage are obtained. The level-by-level splicing method is as follows:

[0129] As shown in FIG. 13, for the (k+1)th level image: The horizontal coordinate of the pixel point in the image is increased by 512, and The vertical coordinate of the pixel point in the image is increased by 512, and the kth level image (m k , n k ) is obtained. Then, the images are spliced level by level, and finally the overall distribution feature image of the tunnel crack damage of the highest level is obtained.

[0130] The overall distribution feature of the tunnel crack damage is the image splicing result after crack recognition by the deep learning model. It not only can be a basis for the subsequent crack spatial distribution feature, and can be directly replaced after the subsequent crack extraction; and in the case of urgent project requirements, the overall distribution feature of the tunnel crack damage can be directly used for kilometer-level tunnel crack spatial distribution model display.

[0131] The embodiment uses the hierarchical storage panoramic development map to store image data with different details at different levels. When processing and analyzing kilometer-level tunnel images in the future, high-level images can be preprocessed to identify areas of interest or perform preliminary analysis, and then these areas can be gradually enlarged to view more details for detailed processing. In this way, the processing efficiency can be improved, and unnecessary consumption of computing resources can be reduced.

[0132] S4, after image binarization of the second bottom layer block, the tunnel crack section crack spatial distribution feature is obtained by expansion and erosion and skeleton extraction;

[0133] Specifically, step S4 specifically includes the following steps:

[0134] S41, according to the boundary of the crack region in the second bottom layer block, the crack region image is extracted using the slicing algorithm in the numpy array;

[0135] Based on the second bottom layer block a′ ij after removing the misrecognized area, the crack region image a′ ij corresponding to the crack region in the second bottom layer block a′ ij-c is extracted, as shown in FIG. 14(b);

[0136] In this embodiment, the numpy array is used to extract the crack site image. First, the coordinate transformation is performed based on the following formula, and the slicing algorithm in the numpy array is used to extract the crack site image, as shown in FIG. 14(b).

[0137] ​​​Wherein: (x0, y0) is the coordinate of the crack area midpoint, x, y is the length and width of the crack area, (x start ,y start ) is the upper left corner coordinate of the crack area, (x end ,y end ) is the right lower corner coordinate of the crack area, and c is the crack area number identified in the bottom block a′ ij .

[0138] S42, the crack area binary image of the crack area image is obtained by an image binarization algorithm;

[0139] According to the crack area image a′ ij-c obtained in S41, the crack damage area binary image b′ ij-c is obtained by using an image binarization algorithm, as shown in Fig. 14.

[0140] Fig. 14(b) is an image before binarization, which is expressed as a three-dimensional array in a numpy array, and each element is a value between 0 and 255. When the three elements of the third dimension are all 0, it means that the pixel point is black. When the three elements of the third dimension are all 255, it means that the pixel point is white.

[0141] Fig. 14(c) is an image after binarization, which is expressed as a two-dimensional array in a numpy array, and each element is 0 or 255. 0 means that the pixel point is black, and 255 means that the pixel point is white.

[0142] Firstly, the image before binarization is converted into a gray image, which is expressed as a two-dimensional array in a numpy array, and each element is a value between 0 and 255. 0 means that the pixel point is black, 255 means that the pixel point is white, and the value between 0 and 255 represents gray. Then, according to the automatically calculated threshold, the elements of the gray image are set to 0 or 255 to obtain the crack area binary image b′ ij-c .

[0143] The data amount of the image after binarization is smaller, and the processing speed is faster, which is beneficial to large-scale image analysis and processing.

[0144] S43, the crack area binary images belonging to the same second bottom block are spliced according to the coordinates to obtain a binary image bottom block;

[0145] The crack area binary images b′ ij-c in the same second bottom block a′ ij are spliced according to the coordinates to obtain the corresponding crack binary image bottom block b′ ij of the bottom block a′ ij after crack damage identification, as shown in Fig. 14(d).

[0146] The specific process is: a blank image of 512*512 pixels is constructed, that is, a numpy two-dimensional array with elements of 255. According to the crack region coordinates obtained in S41 The binary image b' of the crack region obtained in S42 is ij-c The constructed blank image is covered according to the coordinates, as follows:

[0147] S44, the break points of adjacent cracks in the binary image bottom block are connected by the dilation and erosion algorithm, and then the crack region skeleton is extracted by the skeleton extraction algorithm to obtain the crack region skeleton block;

[0148] Specifically, the parameter selection and running times of the dilation and erosion algorithm and the skeleton extraction algorithm in step S44 are different according to the second bottom block crack category.

[0149] The specific process is: first, on the basis of the crack binary image bottom block b' ij , the break points of the identified adjacent cracks are connected by the dilation and erosion algorithm, as shown in FIG. 15(b). Secondly, the crack region skeleton block p' ij is obtained by the skeleton extraction algorithm, as shown in FIG. 15(c).

[0150] Through morphological operations such as dilation and erosion, noise and small interference elements can be removed, and broken cracks can be connected, so that the crack features are more obvious. Then, through skeleton extraction, the basic shape and structure of the crack can be obtained, which is very important for analyzing the spatial distribution characteristics such as the direction, width and length of the crack.

[0151] Due to the different distribution rules of the network cracks and the strip cracks, the parameter selection and running times of the dilation and erosion algorithm and the skeleton extraction algorithm are different in this step. Among them, the network cracks adopt small structure element N times dilation and erosion and skeleton extraction operation, and the strip cracks adopt large structure element single dilation and erosion and skeleton extraction operation.

[0152] In this embodiment, the structure element is only expanded by a single layer of pixels, which is a small structure element, and the structure element is expanded by two layers of pixels or more, which is a large structure element. The value of N is greater than 2, which can be selected according to actual needs, and this place is not limited.

[0153] For example, for network cracks, small structure elements are used for dilation, and then erosion is performed, and so on, after N times, skeleton extraction is performed; for strip cracks, large structure elements are used for dilation, and then erosion and skeleton extraction are performed.

[0154] This embodiment selects different parameters and runs for the expansion corrosion algorithm and the skeleton extraction algorithm according to different crack types, which can preserve the original morphology of the network cracks as much as possible, while ensuring the connectivity of the strip cracks.

[0155] S45. The spatial distribution characteristics of cracks in the tunnel damage section are obtained by splicing the skeleton blocks of the crack area step by step.

[0156] Finally, in the skeletal block p′ of the crack region ij Based on this, the spatial distribution characteristics of cracks in the tunnel damage section are obtained by using the step-by-step splicing method in step S3, as shown in Figure 16.

[0157] S5. Based on the spatial distribution characteristics of cracks in the tunnel's damaged sections and the results of crack regions, construct a kilometer-level tunnel crack spatial distribution and parameter display model.

[0158] Specifically, step S5 includes:

[0159] S51. Calculate the conversion factor between actual distance and pixels: a = d / d p The unit is m / px; d is the length of the area acquired by the cross-sectional image of the inner surface of the designed contour, in meters; d p This is the width of the panoramic unfolded image, in pixels (px).

[0160] S52. Determine the location of the crack in the cross-section and longitudinal section based on the coordinates of the endpoints of the crack skeleton in the crack area skeleton block.

[0161] Based on the tunnel cross-section diagram, the locations of the arch crown (top slab), arch waist (side wall), and arch foot are divided. Combined with the crack skeleton extraction results in the panoramic unfolded diagram, the locations are matched according to the coordinates of the crack skeleton endpoints.

[0162] When the ordinate y of the crack skeleton endpoint satisfies y < y jy1 or y > y jy2 At that time, it indicates that the crack is located at the arch foot.

[0163] When the ordinate y of the crack skeleton endpoint satisfies y jy1 <y<y yd1 or y yd2 <y<y jy2 At that time, it means the crack is located at the arch waist.

[0164] When the ordinate y of the crack skeleton endpoint satisfies y yd1 <y<y yd2 When the crack is located at the top of the arch, it indicates that the crack is located at the top of the arch.

[0165] Where y jy1 With y jy2 The point where the arch foot and the arch waist are separated, y yd1 With y yd2The point where the arch waist and arch crown are separated is obtained based on the cross-sectional design drawing of the crack.

[0166] The x-coordinate of the center point of the crack skeleton represents the location of the crack in the longitudinal section, i.e., the mileage.

[0167] The subsequent repair plan will vary depending on the location of the crack. In this embodiment, the cross-sectional position is used to determine the location of the crack at the top (top plate) and waist (side wall) of the arch; the longitudinal position is used to determine the mileage of the crack. Precise location information helps to develop more targeted and effective repair plans, thereby improving the quality and efficiency of repair.

[0168] S53. Calculate the overall crack length in the image based on the crack area skeleton block, and construct a kilometer-level tunnel crack spatial distribution and crack length display model;

[0169] Specifically, step S53 involves: dividing the crack region skeleton block p′ ij The number of black pixels n′ ij Multiply by the conversion factor 'a' to calculate the crack length l′ of each block. ij ,l′ ij =a·n′ ij Then, by splicing adjacent cracks, the length of the entire crack is obtained, and finally a kilometer-level tunnel crack spatial distribution and crack length display model is constructed.

[0170] Specifically, the construction process of the spatial distribution and crack length display model of the kilometer-level tunnel cracks is as follows: First, a three-dimensional model of the tunnel surface is constructed and the original tunnel image is filled in; second, the crack identification results are drawn. When drawing, the crack length is drawn at the lower left corner of the starting point of each crack, as shown in Figure 17(a), where the unit is m.

[0171] S54. Calculate the average width of the cracks based on the crack area in the second bottom block, and construct a display model of the spatial distribution and average width of cracks in a kilometer-level tunnel.

[0172] Specifically, step S54 is as follows:

[0173] First, calculate the second bottom-level block a′ ij The minimum length d′ of the diagonal of the bounding rectangle of the black pixel ij-c Calculate the crack region a′i j-c Number of black pixels n′ ij-c The ratio method was used to calculate the crack region a′. ij-c The average crack width w′ in ij-c As shown in the following formula:

[0174] w′ ij-c =a·n′ij-c / d′ ij-c wherein c is the second bottom layer block a′ ij identified in the second bottom layer block a′

[0175] Then the average crack width in the block a′ ij identified in the second bottom layer block a′ Finally, a kilometer-level tunnel crack spatial distribution and average crack width display model is constructed.

[0176] Specifically, the kilometer-level tunnel crack spatial distribution and average crack width display model construction process is as follows: first, a three-dimensional model of a tunnel surface is constructed, and the tunnel original image is filled; second, the crack identification result is drawn, and when drawing, the average crack width is drawn at the lower left corner of each crack starting point, as shown in FIG. 17(b), where the unit is mm.

[0177] In the extraction process of the tunnel crack spatial distribution characteristics of the crack section in this embodiment, the binaryzation is used first, so that the image data quantity is smaller and the processing speed is faster, and large-scale image analysis and processing can be effectively performed; and through morphological operations such as dilation and erosion, noise and small interference elements can be removed, and the broken cracks can be connected, so that the crack characteristics are more obvious, and through skeleton extraction, the basic shape and structure of the crack can be obtained, which is more convenient for analyzing the spatial distribution characteristics such as the direction, width and length of the crack, and the efficiency of the crack analysis of the crack section is effectively improved.

[0178] In addition, according to the different parameter selection and operation times of the dilation and erosion algorithm and the skeleton extraction algorithm according to the different crack categories, the original morphology of the network crack can be ensured as much as possible, and the connectivity of the strip crack can be ensured.

[0179] Based on the panoramic development map of the tunnel lining crack section analysis area, and combined with the crack damage depth learning model and the traditional image processing method, the present application can effectively reduce data redundancy, efficiently and accurately realize multi-scale crack identification, and realize crack damage positioning and extraction in a large scale through depth learning, and realize crack damage parameter extraction in a small scale through traditional image processing. The present application can intuitively obtain the overall spatial distribution characteristics and related quantitative parameters of the tunnel lining crack damage in the three-dimensional space on the kilometer level, such as crack length, crack width, crack location, crack cross-section location (such as vault / roof, haunch / side wall, and arch foot), and the like, which provides an important basis for subsequent crack damage specific analysis and has important significance.

[0180] Embodiment Two

[0181] Referring to FIG. 18, the present application discloses a tunnel lining development map crack spatial distribution feature identification device, which comprises the following units:

[0182] A three-dimensional reconstruction unit is configured to acquire a three-dimensional mesh model of a tunnel lining crack section analysis area;

[0183] A panoramic development diagram construction unit is configured to perform panoramic development on the three-dimensional mesh model of the tunnel lining crack section analysis area and a tunnel lining design contour to acquire a panoramic development diagram of the tunnel lining crack section analysis area;

[0184] A crack disease overall distribution feature construction unit is configured to perform crack disease identification on a first bottom layer block of the panoramic development diagram of the tunnel lining crack section analysis area by using a pre-trained crack disease identification model, to obtain a second bottom layer block, and to remove misidentified crack regions according to the distribution of crack region results in the second bottom layer block, and then to perform step-by-step splicing on the second bottom layer block to obtain tunnel crack disease overall distribution features.

[0185] A crack section crack spatial distribution feature construction unit is configured to perform image binarization on the second bottom layer block and then obtain tunnel crack section crack spatial distribution features through inflation corrosion and skeleton extraction.

[0186] A crack region result display unit is configured to construct a kilometer-level tunnel crack spatial distribution and parameter display model according to the tunnel crack section crack spatial distribution features and crack region results.

[0187] The device provided in the embodiments of the present application can execute the crack spatial distribution feature identification method and device based on the tunnel lining development diagram provided in any of the embodiments of the present application, and has the corresponding function modules and beneficial effects of the execution method. It should be noted that the device provided in the embodiments of the present application includes various units and modules only according to functional logic, but is not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for convenient mutual differentiation, and do not limit the protection scope of the present application.

[0188] Embodiment three

[0189] Please refer to FIG. 19, which is a structural schematic diagram of the crack spatial distribution feature identification device based on the tunnel lining development diagram according to the embodiments of the present application. The device 20 includes a processor 21, a memory 22, and a computer program stored in the memory 22 and executable on the processor 21. The processor 21 implements the steps in the above method embodiments when executing the computer program. Alternatively, the processor 21 implements the functions of each module / unit in the above device embodiments when executing the computer program.

[0190] For example, the computer program can be divided into one or more modules / units, which are stored in the memory 22 and executed by the processor 21 to complete the present application. The one or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the tunnel lining unfolding map crack spatial distribution feature identification device 20. For example, the computer program can be divided into the modules in Embodiment Two, and the specific functions of the modules are described in the working process of the device in the above embodiments, which will not be described here.

[0191] The tunnel lining unfolding map crack spatial distribution feature identification device 20 can include, but is not limited to, the processor 21 and the memory 22. Those skilled in the art can understand that the schematic diagram is only an example of the tunnel lining unfolding map crack spatial distribution feature identification device 20, and does not constitute a limitation on the tunnel lining unfolding map crack spatial distribution feature identification device 20, which can include more or fewer components than the diagram, or combine certain components, or different components, for example, the tunnel lining unfolding map crack spatial distribution feature identification device 20 can also include an input / output device, a network access device, a bus, etc.

[0192] The processor 21 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The processor 21 is the control center of the tunnel lining unfolding map crack spatial distribution feature identification device 20, which connects all parts of the tunnel lining unfolding map crack spatial distribution feature identification device 20 through various interfaces and lines.

[0193] The memory 22 can be used to store the computer programs and / or modules, and the processor 21 realizes various functions of the tunnel lining unfolding map crack spatial distribution feature identification device 20 by running or executing the computer programs and / or modules stored in the memory 22, and calling the data stored in the memory 22. The memory 22 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), and the like; and the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), and the like. In addition, the memory 22 can include a high-speed random access memory, and can also include a non-volatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0194] When the modules / units of the tunnel lining unfolding map crack spatial distribution feature identification device 20 are realized in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer-readable storage medium. When the processor 21 executes the computer program, the steps of the above-mentioned various method embodiments can be realized. The computer program includes computer program code, which can be in the form of source code, object code, an executable file, or some intermediate form, etc. The computer-readable medium can include any entity or device capable of carrying the computer program code, a recording medium, a U disk, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the computer-readable medium can include or exclude contents according to the requirements of legislation and patent practice in a jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0195] It should be noted that the apparatus embodiments described above are merely illustrative, and the units described as separate units can or can not be physically separate, and the units displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment of the present application according to actual needs. In addition, the connection relationship between the modules in the apparatus embodiment provided by the present application indicates that there is a communication connection between the modules, which can be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0196] So far, the technical solutions of the present application have been described in conjunction with examples, but those skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application, and the technical solutions after the changes or replacements will fall within the protection scope of the present application.

Claims

1. A method for identifying the spatial distribution characteristics of a tunnel lining unfolding diagram crack, characterized in that, The method comprises the following steps: S1, acquiring a three-dimensional grid model of a tunnel lining crack section analysis area; S2, performing panoramic unfolding on the three-dimensional grid model of the tunnel lining crack section analysis area and a tunnel lining design contour to acquire a panoramic unfolding diagram of the tunnel lining crack section analysis area; S3, using a pre-trained crack damage identification model to identify cracks in a first bottom layer block of the panoramic unfolding diagram of the tunnel lining crack section analysis area, obtaining a second bottom layer block, and removing misidentified crack regions according to the distribution of crack region results in the second bottom layer block, and then sequentially splicing the second bottom layer block to obtain overall distribution characteristics of tunnel crack damage; S4, performing image binarization on the second bottom layer block, and then obtaining crack spatial distribution characteristics of the tunnel crack section through expansion and corrosion and skeleton extraction; S5, constructing a kilometer-level tunnel crack spatial distribution and parameter display model according to the crack spatial distribution characteristics of the tunnel crack section and the crack region results.

2. The method of claim 1, wherein, In step S3, the crack damage identification model is used to identify cracks in the first bottom layer block of the panoramic unfolding diagram of the tunnel lining crack section analysis area, and the second bottom layer block is obtained. The storage mode of the panoramic unfolding diagram of the tunnel lining crack section analysis area is hierarchical storage, including [0, 1,..., N] levels, and each level of image is composed of four low-level images. The Nth level image is acquired as the first bottom layer block. The crack damage identification model is used to identify cracks in the first bottom layer block, and the second bottom layer block containing the identified crack region results is obtained.

3. The method of claim 1, wherein, In step S3, the misidentified crack regions are removed according to the distribution of the crack region results in the second bottom layer block. If there are other crack regions in the judgment region, the crack region is retained. If there are no other crack regions, the crack region is removed. Specifically, all crack regions except the current crack region are traversed, the center point of each remaining crack region is compared with the boundary of the judgment region, and if the center point is within the boundary of the judgment region, it is considered that there are other crack regions in the judgment region, and the region is retained. Otherwise, the current crack region is an isolated region and is removed. Obtain the pixel coordinates of the center point of each crack region in the second bottom layer block, and the number of rows and columns of pixels in the crack region, respectively, and represent them as multiple one-dimensional arrays [l c ,h c The crack region list L is stored in the form of [,dl,dh]. lab In, among which: l c h represents the column coordinates of the pixel point at the center of the crack region. c dl represents the row coordinates of the pixel center point of the crack region, dh represents the number of columns of the crack region, and dh represents the number of rows of the crack region. For each crack region list L lab The list is traversed, and each L lab The list contains all the crack regions; all the crack regions are traversed, and the current crack region is enlarged by a preset determination multiple r as a determination region, and then other crack regions are searched in the determination region, if there are other crack regions in the determination region, the crack region is retained, if not, the crack region is removed.

4. The method of claim 3, wherein, Step S3 further comprises: after obtaining the second bottom layer block, performing gap category identification on the second bottom layer block; based on the crack region center point in the identification result, a a*a pixel square traversal frame is used to traverse the image of the second bottom layer block within the pixel range of each second bottom layer block; if the number of crack region center points in the traversal frame range is greater than a preset crack number threshold f, it is considered that the image is a net-shaped crack; otherwise, it is considered that the image is a strip-shaped crack.

5. The method of claim 1, wherein, Step S4 specifically comprises the following steps:

6. The method of claim 5, wherein, S41, extracting a crack region image from the second bottom layer block using a slicing algorithm in a numpy array according to the boundary of the crack region; S42, obtaining a crack region binarization image of the crack region image through an image binarization algorithm; S43, splicing the crack region binarization images belonging to the same second bottom layer block according to the coordinates to obtain a binarization image bottom layer block; ​ S44, connecting the breakpoints of adjacent cracks in the binarized image bottom layer block by an expansion erosion algorithm, and then extracting a crack region skeleton to obtain a crack region skeleton block by a skeleton extraction algorithm; The parameter selection and operation times of the expansion erosion algorithm and the skeleton extraction algorithm are different according to the second bottom layer block gap category; wherein the small structure element is used for N times expansion erosion and skeleton extraction operation for the net-like crack, and the large structure element is used for single time expansion erosion and skeleton extraction operation for the strip-like crack; the structure element is only expanded by a single layer of pixels for the small structure element, and the structure element is expanded by two layers of pixels or more for the large structure element; S45, the crack region skeleton blocks are spliced level by level to obtain the spatial distribution characteristics of the tunnel crack section.

7. The method of claim 6, wherein, Step S5 specifically includes: S51. Calculate the conversion factor between actual distance and pixels: a = d / d p The unit is m / px; d is the length of the area acquired by the cross-sectional image of the inner surface of the designed contour, in meters; d p This is the width of the panoramic unfolded image, in pixels (px). S52, determining the horizontal and vertical cross-section positions of the cracks according to the crack skeleton endpoint coordinates in the crack region skeleton block: When the longitudinal coordinate y of the crack skeleton end point satisfies y < y jy1 or y > y jy2 , it represents that the crack is located at the springing position, When the longitudinal coordinate y of the crack skeleton end point satisfies y jy1 <y < y yd1 or y yd2 <y < y jy2 , it represents that the crack is located at the haunch position, When the longitudinal coordinate y of the crack skeleton end point satisfies y yd1 <y < y yd2 , it represents that the crack is located at the vault position; where y jy1 where y jy2 is the division point between the haunch and the spandrel, y yd1 where y yd2 is the division point between the haunch and the spandrel, x is the horizontal coordinate of the center point of the crack skeleton, representing the position of the crack in the longitudinal section. S53, calculating the overall crack length in the image according to the crack region skeleton block, and constructing a kilometer-level tunnel crack spatial distribution and crack length display model; S54, calculating the average crack width according to the crack region of the second bottom layer block, and constructing a kilometer-level tunnel crack spatial distribution and average crack width display model.

8. The method of claim 7, wherein, Step S53 specifically involves: dividing the crack region skeleton block p′ ij The number of black pixels n′ ij Multiply by the conversion factor 'a' to calculate the crack length l′ of each block. ij Then, by splicing adjacent cracks, the length of the entire crack is obtained, and finally a kilometer-level tunnel crack spatial distribution and crack length display model is constructed.

9. The method of claim 7, wherein, The step S54 specifically includes: First, the second bottom block a' is calculated ij The diagonal length d' of the minimum circumscribed rectangle of the black pixel points in the middle ij-c The crack area a' is calculated ij-c The number of black pixel points n' in the middle ij-c The average crack width w' in the crack area a' is calculated by the ratio method ij-c ij-c As shown in the following formula:​ w′ ij-c = a · n′ ij-c / d′ ij-c where c is the crack region number identified within the second base layer tile a′ ij . The average crack width in the zone a' is then calculated according to the following formula ij ​ Finally, a kilometer-level tunnel crack spatial distribution and average crack width display model is constructed.

10. A device for identifying the spatial distribution of damage based on a tunnel lining unfolding map, characterized in that It includes the following units: A three-dimensional reconstruction unit for obtaining a three-dimensional grid model of a tunnel lining crack section analysis area; A panoramic development diagram construction unit for developing a three-dimensional grid model of a tunnel lining crack section analysis area and a tunnel lining design contour to obtain a panoramic development diagram of a tunnel lining crack section analysis area; A crack damage overall distribution feature construction unit that uses a pre-trained crack damage identification model to identify the crack damage of a first bottom layer block of the panoramic development diagram of the tunnel lining crack section analysis area to obtain a second bottom layer block, and removes misidentified crack regions according to the distribution of the crack region results in the second bottom layer block, and then splices the second bottom layer block level by level to obtain the overall distribution characteristics of the tunnel crack damage; A crack section crack spatial distribution feature construction unit for binarizing the second bottom layer block and then obtaining the tunnel crack section crack spatial distribution characteristics by expansion erosion and skeleton extraction; A crack region result display unit for constructing a kilometer-level tunnel crack spatial distribution and parameter display model according to the tunnel crack section crack spatial distribution characteristics and the crack region results.

Images