Tunnel lining surface crack risk detection method and system and medium

Abstract

The invention provides a tunnel lining surface crack risk detection method and system and a medium, and relates to the technical field of crack detection, two groups of identification images are generated through image acquisition and processing, one group is graying processing, and the other group is HSV color space conversion. Then, gray region blocks are extracted from the first recognition image by using canny edge detection, color region blocks in one-to-one correspondence with the gray region blocks are segmented from the second recognition image, HSV color characteristics and edge pixel point distribution are extracted respectively, and the hue mean value, the brightness and the crack strength of the region blocks are calculated; risk indexes of the region blocks are generated based on the data, and the comprehensive risk index of the tunnel lining surface is further calculated. And finally, generating a high-risk threshold value and a low-risk threshold value, and judging the severity of the crack risk by comparing the relationship between the real-time comprehensive risk index and the threshold values.

Description

Technical Field

[0001] The present invention relates to the technical field of crack detection, and specifically to a method, system and medium for detecting the risk of cracks on the surface of tunnel lining. Background Art

[0002] In tunnel construction and maintenance, the tunnel lining is a key part to ensure the safety and stability of the tunnel. Its integrity is directly related to the service life of the tunnel and the driving safety. However, cracks often occur on the surface of the tunnel lining due to factors such as geological conditions, construction quality and service life. If not treated, it may lead to structural damage and even safety accidents. The crack detection on the lining surface is a common and difficult technical problem. Traditional detection methods mostly rely on manual observation or physical detection, which are not only time-consuming and laborious, but also often difficult to achieve real-time monitoring of cracks. Although the existing crack identification through image processing technology can identify cracks to a certain extent, most methods lack systematicness, the feature extraction and analysis of cracks are not fine enough, it is easy to miss detection or misdetection, and a comprehensive risk assessment cannot be provided. Therefore, there is an urgent need for an efficient and real-time method for detecting cracks on the surface of tunnel lining to improve the detection efficiency and accuracy and ensure the safe operation of the tunnel.

[0003] In the prior art, the publication number CN118864447A discloses a method, system, device and medium for detecting cracks on the surface of tunnel lining. Each image in the sample training set is subjected to multiple feature extractions. The specific feature extraction processes include horizontal feature processing, vertical feature processing and average feature processing, and the feature maps obtained from multiple feature extractions are fused to obtain a feature fusion map corresponding to each image in the sample training set; each feature fusion map is respectively subjected to feature optimization processing, and the feature maps after the feature optimization processing are input into a preset initial model for processing until the initial model reaches the preset training end condition to generate a crack detection model. The image to be detected is input into the crack detection model to obtain the crack detection result.

[0004] The main problems of the above solution are: Although multiple feature extractions can be used to generate a feature fusion map, which can extract various features, the complexity and computational amount are relatively high, the processing efficiency is low, and it is difficult to apply to large-scale real-time monitoring scenarios; the generated crack detection model shows poor generalization ability in practical applications, and the detection accuracy decreases when detecting different types or states of tunnel lining images.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure. Therefore, it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The purpose of the present invention is to provide a method, system and medium for detecting the risk of surface cracks in tunnel linings, so as to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for detecting the risk of surface cracks in tunnel linings, the specific steps include:

[0009] Step 1: Collect images of the tunnel lining surface, scale the images to a size of 224×224, and copy them into two identical groups. One group is grayscale processed to generate a first recognition image, and the other group is converted from the RGB color space to the HSV color space to generate a second recognition image, and the pixel positions of the first recognition image and the second recognition image are mapped one by one.

[0010] Step 2: Based on the canny edge detection technology, extract features from the first recognition image to obtain multiple grayscale region blocks, and each region block corresponds to a crack region.

[0011] Step 3: Segment the corresponding color region blocks from the second recognition image, generate the hue mean and brightness of the region blocks based on the characteristics of the HSV color space, generate the color uniformity of the region blocks based on the hue mean of each region block, and generate the crack intensity of the region blocks based on the number of edge pixels and the pixels in the region surrounded by the edge pixels.

[0012] Step 4: Generate the risk index of each region block based on the crack intensity and brightness of each region block, and generate the comprehensive risk index of the collected tunnel lining surface based on the risk index of each region block and the color uniformity of the region block.

[0013] Step 5: Obtain the comprehensive risk indices of the tunnel lining surfaces that are known to be defined as high-risk, medium-risk and low-risk cracks, generate a high-risk threshold based on the lowest comprehensive risk index of high-risk cracks and the highest comprehensive risk index of medium-risk cracks, generate a low-risk threshold based on the lowest comprehensive risk index of medium-risk cracks and the highest comprehensive risk index of low-risk cracks, and judge the severity of the crack risk according to the size relationship between the real-time generated comprehensive risk index and the high-risk threshold and the low-risk threshold.

[0014] Further, establish a plane rectangular coordinate system with the column where the leftmost pixel point of the first recognition image is located as the y-axis and the row where the bottommost pixel point is located as the x-axis, and map each pixel point coordinate to the second recognition image to ensure that the same pixel point in the two images has a unique coordinate.

[0015] Further, the principle for extracting the grayscale region blocks of the first recognition image is:

[0016] For each pixel point in the first recognition image, the pixel point and its neighboring pixel points are respectively convolved with the horizontal direction template and the vertical direction template of the Prewitt operator to generate the gray - level differences of the pixel point in the horizontal and vertical directions. The formula is as follows:

[0017]

[0018]

[0019] Among them, P X represents the horizontal direction template of the Prewitt operator, P Y represents the vertical direction template of the Prewitt operator, G x represents the horizontal direction difference of the pixel point, G y represents the vertical direction difference of the pixel point, and (x, y) represents the coordinates of the pixel point;

[0020] According to the gray - level differences in the horizontal and vertical directions, the gradient magnitude of each pixel point is generated. The formula is as follows:

[0021]

[0022] Among them, G(x, y) represents the gradient magnitude of the pixel point with coordinates (x, y), G x represents the horizontal direction difference of the pixel point, G y represents the vertical direction difference of the pixel point;

[0023] A preset edge threshold is set. When the gradient magnitude of a pixel point is higher than the edge threshold, it is regarded as an edge pixel point. Based on the edge pixel points, multiple gray - level region blocks are segmented from the first recognition image.

[0024] Furthermore, the hue mean and brightness of the region block are generated, and the color uniformity of the region block is generated based on the hue mean of each region block. The principle is as follows:

[0025] The formula for generating the hue mean of the region block is as follows:

[0026]

[0027] Among them, H m represents the hue mean of the m - th region block, m represents the index of the region block, R m represents the number of pixel points in the m - th region block, h i (m) represents the hue of the i - th pixel point in the m - th region block, and i represents the index of the pixel point in the region block;

[0028] The formula for generating the brightness of the region block is as follows:

[0029]

[0030] Among them, L m represents the brightness of the m-th region block, and v i (m) represents the brightness of the i-th pixel in the m-th region block;

[0031] The formula for generating the color uniformity of the region block is:

[0032]

[0033] Among them, μ H represents the average of the hue means of all region blocks, M represents the number of region blocks, σ H represents the standard deviation of the hue means of the region blocks, C represents the color uniformity, H max represents the maximum hue mean of the region blocks, H min represents the minimum hue mean of the region blocks.

[0034] Furthermore, the formula for generating the crack intensity of the region block is:

[0035]

[0036] Among them, S m represents the crack intensity of the m-th region block, R m,edge represents the number of edge pixels of the m-th region block, R m,in represents the number of pixels in the region surrounded by the edge pixels in the m-th region block, and α, β, γ represent weight coefficients, and α + β + γ = 1.

[0037] Furthermore, the principle for generating the comprehensive risk index is:

[0038] The formula for generating the risk index of the region block is:

[0039]

[0040] Among them, F m represents the risk index of the m-th region block;

[0041] The formula for generating the comprehensive risk index is:

[0042]

[0043] Among them, represents the average risk index of the region block, and K represents the comprehensive risk index.

[0044] Furthermore, the principle for generating the high-risk threshold and the low-risk threshold is:

[0045]

[0046] Among them, T high represents the high-risk threshold, U mid represents the highest comprehensive risk index of medium-risk cracks, D high represents the lowest comprehensive risk index of high-risk cracks, U low represents the highest risk index of low-risk cracks, D mid represents the lowest comprehensive risk index of medium-risk cracks;

[0047] When K ≥ T high it is determined as a high-risk crack;

[0048] When T low ≤ K < T high it is determined as a medium-risk crack;

[0049] When K < T low it is determined as a low-risk crack.

[0050] The present invention also provides a risk detection system for cracks on the surface of a tunnel lining. The system is used to implement the above-mentioned risk detection method for cracks on the surface of a tunnel lining, and specifically includes:

[0051] An image acquisition module, which is used to acquire an image of the surface of the tunnel lining, scale the image to a size of 224×224, and copy it into two identical groups. One group is grayscale processed to generate a first recognition image, and the other group converts the image from the RGB color space to the HSV color space to generate a second recognition image, and maps the pixel positions of the first recognition image and the second recognition image one by one;

[0052] An edge extraction module, which is used to extract features from the first recognition image based on the canny edge detection technology to obtain a plurality of grayscale region blocks, and each region block corresponds to a crack region;

[0053] A feature extraction module, which splits out color region blocks corresponding one by one to the grayscale region blocks from the second recognition image, generates the hue mean and brightness of the region blocks based on the characteristics of the HSV color space, generates the color uniformity of the region blocks based on the hue mean of each region block, and generates the crack intensity of the region blocks based on the number of edge pixel points and the pixel points in the region surrounded by the edge pixel points;

[0054] A risk calculation module, which is used to generate the risk index of the region block based on the crack intensity and brightness of each region block, and generate the comprehensive risk index of the surface of the tunnel lining collected based on the risk index of each region block and the color uniformity of the region block;

[0055] A risk assessment module is used to obtain the comprehensive risk indices of the tunnel lining surface for known cracks defined as high-risk, medium-risk, and low-risk cracks, generate a high-risk threshold based on the lowest comprehensive risk index of high-risk cracks and the highest comprehensive risk index of medium-risk cracks, generate a low-risk threshold based on the lowest comprehensive risk index of medium-risk cracks and the highest comprehensive risk index of low-risk cracks, and determine the severity of the crack risk according to the magnitude relationship between the real-time generated comprehensive risk index and the high-risk threshold and the low-risk threshold.

[0056] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed on a computer, the computer is made to execute the above-mentioned method for detecting the risk of cracks on the tunnel lining surface.

[0057] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0058] The present invention divides the tunnel lining image into a grayscale image and an HSV color space image, determines the edge pixel points in the grayscale image, and determines the hue of the pixel points in the HSV color space image, realizing multi-dimensional information fusion, which can more comprehensively reflect the actual situation of the lining surface and does not solely rely on a single feature, thereby improving the reliability of crack detection; divides the overall image into several identical regional blocks, generates a risk index for each regional block, enabling the colors and crack characteristics in different regions to be independently identified, not only improving the speed of risk assessment but also enhancing the accuracy of risk detection.

[0059] The present invention also combines crack intensity, brightness, and color uniformity to more precisely identify potential high-risk areas and capture early signs of crack occurrence, thereby providing a basis for preventive maintenance; generates high-risk and low-risk thresholds through the average value and standard deviation of historical comprehensive risk indices, ensuring the scientificity and rationality of threshold setting. By comparing with the high-risk threshold and the low-risk threshold, the severity of the crack risk can be clearly determined. This systematic method provides a clear risk level for decision-makers, enabling the formulation of corresponding maintenance strategies and the adoption of necessary safety measures while identifying cracks. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a schematic flowchart of the method according to an embodiment of the present invention;

[0061] Figure 2 It is a schematic diagram of the system module according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0062] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to specific embodiments.

[0063] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention shall have the ordinary meanings understood by those of ordinary skill in the field to which the present invention pertains. The "first", "second" and similar terms used in the present invention do not denote any order, quantity or importance, but are only used to distinguish different components. Words such as "comprising" or "including" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. Words such as "connected" or "linked" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Upper", "lower", "left", "right", etc. are only used to represent relative position relationships, and when the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0064] Embodiment:

[0065] Please refer to Figure 1 , the present invention provides a technical solution:

[0066] A method for detecting the risk of surface cracks in tunnel linings, the specific steps include:

[0067] Step 1: Collect images of the surface of the tunnel lining, scale the images to a size of 224×224, and copy them into two identical groups. One group is subjected to grayscale processing to generate a first recognition image, and the other group converts the image from the RGB color space into the HSV color space to generate a second recognition image, and maps the pixel positions of the first recognition image and the second recognition image one by one;

[0068] In this embodiment, a plane rectangular coordinate system is established with the column where the leftmost pixel point of the first recognition image is located as the y-axis and the row where the lowermost pixel point is located as the x-axis, and the coordinates of each pixel point are mapped into the second recognition image to ensure that the same pixel point in the two images has a unique coordinate.

[0069] The principle for performing grayscale processing is:

[0070] Y = 0.299·R + 0.587·G + 0.114·B

[0071] where Y represents the grayscale value, R represents the red channel value, G represents the green channel value, and B represents the blue channel value

[0072] The principle for converting the image from the RGB color space into the HSV color space is:

[0073] For any point in the image, its RGB color space is (R, G, B), and its HSV color space is (H, S, V). First, the values of R, G, and B are converted to the range of [0, 1]:

[0074]

[0075]

[0076] Generate H, S, and V based on R0, G0, and B0, and the formulas used are as follows:

[0077] V = max(R0, G0, B0)

[0078]

[0079] Among them, H, S, and V represent hue, saturation, and brightness respectively.

[0080] Step 2: Extract features from the first recognition image based on the canny edge detection technique to obtain multiple grayscale region blocks, and each region block corresponds to a crack region;

[0081] In this embodiment, the principle for extracting the grayscale region blocks of the first recognition image is as follows:

[0082] For each pixel point in the first recognition image, convolve the pixel point and its neighborhood pixel points with the horizontal direction template and vertical direction template of the Prewitt operator respectively to generate the grayscale differences of the pixel point in the horizontal and vertical directions. The formulas used are as follows:

[0083]

[0084] Among them, P X represents the horizontal direction template of the Prewitt operator, P Y represents the vertical direction template of the Prewitt operator, G x represents the horizontal direction difference of the pixel point, G y represents the vertical direction difference of the pixel point, and (x, y) represents the coordinates of the pixel point;

[0085] Generate the gradient magnitude of each pixel point according to the grayscale differences in the horizontal and vertical directions. The formula used is as follows:

[0086]

[0087] Among them, G(x, y) represents the gradient magnitude of the pixel point with coordinates (x, y), G x represents the horizontal direction difference of the pixel point, G y represents the vertical direction difference of the pixel point;

[0088] Preset an edge threshold. When the gradient magnitude of a pixel is higher than the edge threshold, it is regarded as an edge pixel. Based on the edge pixels, multiple grayscale region blocks are segmented from the first recognition image.

[0089] The gradient magnitude reflects the rate of brightness change in the image. Edge points are usually located where the brightness changes drastically in the image. By calculating the gradient magnitude of each pixel to identify these changes, the more drastic the brightness change, the more likely it is an edge pixel. Therefore, an edge threshold is set to screen out the pixels with a gradient magnitude greater than the edge threshold as edge pixels.

[0090] Step 3: Segment color region blocks corresponding one by one to the grayscale region blocks from the second recognition image, generate the hue mean and brightness of the region blocks based on the HSV color space characteristics, generate the color uniformity of the region blocks based on the hue mean of each region block, and generate the crack intensity of the region blocks based on the edge pixels and the number of pixels in the area surrounded by the edge pixels.

[0091] In this embodiment, the principle for generating the hue mean and brightness of the region blocks and generating the color uniformity of the region blocks based on the hue mean of each region block is as follows:

[0092] The formula for generating the hue mean of the region block is:

[0093]

[0094] where H m represents the hue mean of the m-th region block, m represents the index of the region block, R m represents the number of pixels in the m-th region block, h i (m) represents the hue of the i-th pixel in the m-th region block, and i represents the index of the pixel in the region block;

[0095] The hue mean of the entire region block is generated by the average hue of the pixels within the region block;

[0096] The formula for generating the brightness of the region block is:

[0097]

[0098] where L m represents the brightness of the m-th region block, v i (m) represents the brightness of the i-th pixel in the m-th region block;

[0099] The brightness of the entire region block is generated by the average brightness of all the pixels within the region block;

[0100] The formula for generating the color uniformity of the region block is:

[0101]

[0102] Among them, μ H represents the average of the hue means of all region blocks, M represents the number of region blocks, and σ H represents the standard deviation of the hue means of the region blocks, C represents the color uniformity, and H max represents the maximum hue mean of the region blocks, and H min represents the minimum hue mean of the region blocks.

[0103] When cracks appear on the surface of the tunnel lining, the color of the crack area will be different from that of the surrounding area, resulting in uneven overall color distribution, that is, the color uniformity decreases. The color uniformity reflects the severity of the crack from the level of hue mean, and the standard deviation reflects the degree of dispersion of the hue means of the region blocks. The larger the standard deviation, the greater the fluctuation of the hue mean and the lower the color uniformity. H max -H min represents the range of the hue means within the image, which is used to normalize the standard deviation. Dividing the standard deviation by the range can, while normalizing, intuitively understand the distribution of the hue means within the overall hue mean range. The color uniformity is inversely proportional to the standard deviation of the hue means.

[0104] The formula for generating the crack intensity of the region block is:

[0105]

[0106] Among them, S m represents the crack intensity of the m-th region block, R m,edge represents the number of edge pixel points of the m-th region block, and R m,in represents the number of pixel points in the region surrounded by the edge pixel points of the m-th region block. α, β, and γ represent weight coefficients, and α + β + γ = 1.

[0107] The crack intensity reflects the crack risk of a region block. The greater the crack intensity, the greater the crack risk. R m,edge represents the number of edge pixel points. The more edge pixel points, the larger the crack. is the ratio of the edge pixel points to all pixel points in the region block, which reflects the density of the edge. The higher the density, the more concentrated and numerous the cracks per unit area. R m,in represents the number of pixel points surrounded by the edge pixel points, which reflects the size and extent of the crack. The larger R m,in , the more likely the crack is to extend. The crack intensity of the region block is proportional to R m,edge , and R m,in ; R m,edge and directly reflects the size and quantity of the cracks, so it has a higher weight, R m,in mainly reflects the extension trend of the cracks, so it has a lower weight. The specific values of the weight coefficients are: α = 0.4, β = 0.4, γ = 0.2.

[0108] Step 4: Generate the risk index of each regional block based on the crack intensity and brightness of the regional block, and generate the comprehensive risk index of the collected tunnel lining surface based on the risk index of each regional block and the color uniformity of the regional block;

[0109] In this embodiment, the principle for generating the comprehensive risk index is:

[0110] The formula for generating the risk index of the regional block is:

[0111]

[0112] where, F m represents the risk index of the m-th regional block;

[0113] The greater the crack intensity, it indicates that the possible crack area of the regional block is large and the number of cracks is large. Therefore, the risk index of the regional block rises. At the same time, due to the appearance of cracks, the light illumination on the surface of the regional block is blocked and the brightness decreases. Under the same light intensity, the more and deeper the cracks, the lower the brightness of the regional block. Therefore, the risk index of the regional block is directly proportional to the crack intensity and inversely proportional to the brightness of the regional block.

[0114] The formula for generating the comprehensive risk index is:

[0115]

[0116] where, represents the average risk index of the regional block, and K represents the comprehensive risk index.

[0117] The comprehensive risk index reflects the crack risk situation of the entire image. The average risk index of the regional block is generated based on the risk index of each regional block. The higher the average risk index, it indicates that the overall risk of the image is greater. The higher the color uniformity of the image, it indicates that the color change of the image is concentrated in a smaller range. The comprehensive risk index is directly proportional to the average risk index of the regional block and inversely proportional to the color uniformity of the image. At the same time, through the exponential form of e, it can better reflect the cumulative effect of the regional block risk, making the change of the data more obvious and facilitating the timely discovery of risks.

[0118] Step 5: Obtain the comprehensive risk index of the tunnel lining surface that is defined as high-risk, medium-risk, and low-risk cracks. Generate a high-risk threshold based on the lowest comprehensive risk index of high-risk cracks and the highest comprehensive risk index of medium-risk cracks, and generate a low-risk threshold based on the lowest comprehensive risk index of medium-risk cracks and the highest comprehensive risk index of low-risk cracks. Determine the severity of the crack risk according to the magnitude relationship between the real-time generated comprehensive risk index and the high-risk threshold and the low-risk threshold.

[0119] In this embodiment, the principles for generating the high-risk threshold and the low-risk threshold are as follows:

[0120]

[0121] Among them, T high represents the high-risk threshold, U mid represents the highest comprehensive risk index of medium-risk cracks, D high represents the lowest comprehensive risk index of high-risk cracks, U low represents the highest risk index of low-risk cracks, D mid represents the lowest comprehensive risk index of medium-risk cracks;

[0122] When K ≥ T high , it is determined as a high-risk crack;

[0123] When T low ≤ K < T high , it is determined as a medium-risk crack;

[0124] When K < T low , it is determined as a low-risk crack.

[0125] Collect a large number of cracks with known risk levels, calculate the comprehensive risk index of each crack. The comprehensive risk indices of high-risk cracks, medium-risk cracks, and low-risk cracks are respectively within a certain range. Take the average value of the lowest risk index of high-risk cracks and the highest risk index of medium-risk cracks as the high-risk threshold, and the average value of the lowest risk index of medium-risk cracks and the highest risk index of low-risk cracks as the low-risk threshold.

[0126] Please refer to Figure 2 , the present invention also provides a tunnel lining surface crack risk detection system, which is used to implement the above-mentioned tunnel lining surface crack risk detection method, and specifically includes:

[0127] The image acquisition module is used to acquire the images of the tunnel lining surface, scale the images to a size of 224×224, and copy them into two identical groups. One group is grayscaled to generate the first recognition image, and the other group is transformed from the RGB color space to the HSV color space to generate the second recognition image, and the pixel positions of the first recognition image and the second recognition image are mapped one by one;

[0128] The edge extraction module is used to extract features from the first recognition image based on the canny edge detection technology to obtain multiple grayscale region blocks, and each region block corresponds to a crack region;

[0129] The feature extraction module segments the corresponding color region blocks from the second recognition image one by one with the grayscale region blocks, generates the hue mean and brightness of the region blocks based on the HSV color space characteristics, generates the color uniformity of the region blocks based on the hue mean of each region block, and generates the crack intensity of the region blocks based on the number of edge pixels and the pixels in the region surrounded by the edge pixels;

[0130] The risk calculation module is used to generate the risk index of the region block based on the crack intensity and brightness of each region block, and generate the comprehensive risk index of the acquired tunnel lining surface based on the risk index of each region block and the color uniformity of the region block;

[0131] The risk assessment module is used to obtain the comprehensive risk indices of the tunnel lining surface that are known to be defined as high-risk, medium-risk, and low-risk cracks, generate the high-risk threshold based on the lowest comprehensive risk index of the high-risk cracks and the highest comprehensive risk index of the medium-risk cracks, generate the low-risk threshold based on the lowest comprehensive risk index of the medium-risk cracks and the highest comprehensive risk index of the low-risk cracks, and judge the severity of the crack risk according to the size relationship between the real-time generated comprehensive risk index and the high-risk threshold and the low-risk threshold.

[0132] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed in a computer, the computer is made to execute the above-mentioned method for detecting the crack risk of the tunnel lining surface.

[0133] All the above formulas are dimensionless and take their numerical values for calculation. The formulas are obtained by software simulation of collecting a large amount of data to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0134] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will realize that the units and algorithm steps of each example described in connection with the embodiments disclosed herein can be implemented by electronic hardware, or by a combination of computer software and electronic hardware. Whether these functions are executed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0135] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units. They may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0136] As described above, the specific implementation manners of the present application are only described, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present application, and all of them should be covered by the protection scope of the present application.

Claims

1. A method for detecting the risk of surface cracks in tunnel linings, characterized in that, The specific steps include: Step 1: Collect images of the tunnel lining surface, scale the images to a size of 224×224, and copy them into two identical groups. One group is grayscaled to generate a first recognition image, and the other group is converted from the RGB color space to the HSV color space to generate a second recognition image. Then, the pixel positions of the first recognition image and the second recognition image are mapped one by one. Step 2: Based on the canny edge detection technology, extract features from the first recognition image to obtain multiple grayscale region blocks, and each region block corresponds to a crack region. Step 3: Segment the corresponding color region blocks from the second recognition image. Generate the hue mean and brightness of the region blocks based on the characteristics of the HSV color space, generate the color uniformity of the region blocks based on the hue mean of each region block, and generate the crack intensity of the region blocks based on the edge pixel points and the number of pixel points in the region surrounded by the edge pixel points. Step 4: Generate the risk index of the region block based on the crack intensity and brightness of each region block, and generate the comprehensive risk index of the collected tunnel lining surface based on the risk index of each region block and the color uniformity of the region block. Step 5: Obtain the comprehensive risk indices of the tunnel lining surfaces that are known to be defined as high-risk, medium-risk, and low-risk cracks. Generate a high-risk threshold based on the lowest comprehensive risk index of the high-risk cracks and the highest comprehensive risk index of the medium-risk cracks, generate a low-risk threshold based on the lowest comprehensive risk index of the medium-risk cracks and the highest comprehensive risk index of the low-risk cracks, and judge the severity of the crack risk according to the size relationship between the real-time generated comprehensive risk index and the high-risk threshold and the low-risk threshold.

2. The method for detecting the risk of surface cracks in a tunnel lining according to claim 1, wherein: In step 1, a plane rectangular coordinate system is established with the column where the leftmost pixel point of the first recognition image is located as the y-axis and the row where the bottommost pixel point is located as the x-axis, and the coordinates of each pixel point are mapped into the second recognition image to ensure that the same pixel point in the two images has a unique coordinate.

3. A method for detecting the risk of surface cracks in a tunnel lining according to claim 1, characterized in that: The principle for extracting the grayscale region blocks of the first recognition image in step 2 is: For each pixel point in the first recognition image, the pixel point and its neighborhood pixel points are respectively convolved with the horizontal direction template and the vertical direction template of the Prewitt operator to generate the grayscale difference of the pixel point in the horizontal direction and the vertical direction. The formula is: Among them, P X represents the horizontal direction template of the Prewitt operator, P Y represents the vertical direction template of the Prewitt operator, G x represents the horizontal direction difference of the pixel point, G y represents the vertical direction difference of the pixel point, and (x, y) represents the coordinates of the pixel point; According to the grayscale differences in the horizontal direction and the vertical direction, generate the gradient amplitude of each pixel point. The formula is: Among them, G(x, y) represents the gradient magnitude of the pixel point with coordinates (x, y), and G x represents the horizontal difference of the pixel point, and G y represents the vertical difference of the pixel point; Preset an edge threshold. When the gradient amplitude of a pixel point is higher than the edge threshold, it is used as an edge pixel point, and multiple grayscale region blocks are segmented from the first recognition image based on the edge pixel points.

4. A method for detecting the risk of surface cracks in a tunnel lining according to claim 1, characterized in that: The principle for generating the hue mean and brightness of the region block and generating the color uniformity of the region block based on the hue mean of each region block is: The formula for generating the hue mean of the region block is: Among them, H m represents the hue mean of the m-th regional block, where m represents the index of the regional block, and R m represents the number of pixels in the m-th regional block, and h i (m) represents the hue of the i-th pixel in the m-th regional block, where i represents the index of the pixel in the regional block; The formula for generating the brightness of the region block is: Among them, L m represents the luminance of the m-th region block, and v i (m) represents the luminance of the i-th pixel in the m-th region block; The formula for generating the color uniformity of the region block is: Among them, μ H represents the average of the hue means of all region blocks, M represents the number of region blocks, σ H represents the standard deviation of the hue means of the region blocks, C represents the color uniformity, H max represents the maximum hue mean of the region blocks, H min represents the minimum hue mean of the region blocks.

5. A method for detecting the risk of surface cracks in a tunnel lining according to claim 4, characterized in that: The formula for generating the crack intensity of the region block is: Among them, S m represents the crack intensity of the m-th regional block, R m,edge represents the number of edge pixel points of the m-th regional block, R m,in represents the number of pixel points in the area surrounded by edge pixel points in the m-th regional block, and α, β, and γ represent weight coefficients, and α + β + γ = 1.

6. The method for detecting the risk of surface cracks in a tunnel lining according to claim 4, characterized in that: The principle for generating the comprehensive risk index is: The formula for generating the risk index of the region block is: Among them, F m represents the risk index of the m-th regional block; The formula for generating the comprehensive risk index is: Among them, represents the average risk index of the regional block, and K represents the comprehensive risk index.

7. A method for detecting the risk of surface cracks in tunnel linings according to claim 6, characterized in that: The principle for generating the high-risk threshold and the low-risk threshold in step 5 is as follows: Among them, T high represents the high-risk threshold, U mid represents the highest comprehensive risk index of medium-risk cracks, D high represents the lowest comprehensive risk index of high-risk cracks, U low represents the highest risk index of low-risk cracks, D mid represents the lowest comprehensive risk index of medium-risk cracks; When K ≥ T high it is judged as a high-risk crack; When T low ≤K<T high it is judged as a medium-risk crack; When K < T low it is judged as a low-risk crack.

8. A risk detection system for surface cracks of tunnel lining, characterized in that: The system is used to implement the tunnel lining surface crack risk detection method described in any one of claims 1-7, and specifically includes: An image acquisition module, which is used to acquire the tunnel lining surface image, scale the image to a size of 224×224, and copy it into two identical groups. One group is grayscaled to generate the first recognition image, and the other group is converted from the RGB color space to the HSV color space to generate the second recognition image, and the pixel positions of the first recognition image and the second recognition image are mapped one by one; An edge extraction module, which is used to extract features from the first recognition image based on the canny edge detection technology to obtain multiple gray-scale region blocks, and each region block corresponds to a crack region; A feature extraction module, which segments the color region blocks corresponding one by one to the gray-scale region blocks from the second recognition image, generates the hue mean and brightness of the region blocks based on the HSV color space characteristics, generates the color uniformity of the region blocks based on the hue mean of each region block, and generates the crack intensity of the region blocks based on the number of edge pixel points and the pixel points in the region surrounded by the edge pixel points; A risk calculation module, which is used to generate the risk index of the region block based on the crack intensity and brightness of each region block, and generate the comprehensive risk index of the acquired tunnel lining surface based on the risk index of each region block and the color uniformity of the region block; A risk assessment module, which is used to obtain the comprehensive risk indices of the tunnel lining surface known to be defined as high-risk, medium-risk, and low-risk cracks, generate the high-risk threshold based on the lowest comprehensive risk index of the high-risk cracks and the highest comprehensive risk index of the medium-risk cracks, generate the low-risk threshold based on the lowest comprehensive risk index of the medium-risk cracks and the highest comprehensive risk index of the low-risk cracks, and judge the severity of the crack risk according to the size relationship between the real-time generated comprehensive risk index and the high-risk threshold and the low-risk threshold.

9. A computer-readable storage medium, characterized in that: It stores a computer program, and when the computer program is executed in a computer, it causes the computer to execute the tunnel lining surface crack risk detection method described in any one of claims 1-7.

Images