Bridge surface crack detection method and system based on image data processing

Abstract

The invention relates to the technical field of bridge surface crack detection, and particularly discloses a bridge surface crack detection method and system based on image data processing, and the method comprises the steps: analyzing the basic information of a bridge, matching a bridge detection camera deployment scheme, obtaining the quality data of each image shot by a bridge detection camera, and obtaining the quality data of each image; judging whether the image quality is qualified or not, if the image quality is qualified, extracting each suspected crack area, analyzing each suspected crack area, judging whether the suspected crack area is a bridge surface crack or not, marking the crack area, analyzing the bridge surface crack state, and determining a bridge surface crack detection grade result; according to the method, the problems that traditional bridge surface crack detection is insufficient in dynamic adaptive capacity to the bridge, personalized detection for different bridge surface cracks cannot be realized, and the manual detection efficiency is low are solved, the cracks can be accurately detected, and the state of the bridge surface cracks can be analyzed according to the judgment result of each suspected crack area.

Description

Technical Field

[0001] The present invention relates to the technical field of bridge surface crack detection, and specifically to a method and system for bridge surface crack detection based on image data processing. Background Art

[0002] Bridges are a key component of transportation infrastructure and play a crucial role in economic development and social interaction. Whether it is a highway bridge, a railway bridge, or a pedestrian overpass in the city, they carry a large amount of vehicle and pedestrian traffic. Once there are problems with the bridge structure, especially diseases such as surface cracks, it may lead to a decrease in the bearing capacity of the bridge, a reduction in durability, and even may cause the collapse of the bridge, resulting in serious casualties and losses. With the continuous development of digital image technology, the performance of camera equipment has been greatly improved, and it can obtain high-resolution and high-quality images of the bridge surface. At the same time, the enhancement of computer processing power makes it possible to process a large amount of image data. It can achieve large-area and high-precision detection of the bridge surface without damaging the bridge structure, providing more scientific and accurate data support for the maintenance and management of the bridge.

[0003] Nowadays, there are still some deficiencies in the research on bridge surface crack detection based on image data processing. Specifically, traditional bridge surface crack detection mainly relies on manual detection, that is, the detection personnel check the bridge surface by visual observation or using simple tools. This method not only has low efficiency, but also is greatly affected by the subjective factors of the detection personnel. In addition, for some difficult-to-reach areas (such as the high piers of the bridge, the bottom of the beam body, etc.), manual detection may have safety risks and it is difficult to conduct a comprehensive and detailed inspection. Moreover, traditional bridge surface crack detection has insufficient dynamic adaptability to the bridge, and it is impossible to achieve personalized detection for different bridge surface cracks, resulting in inaccurate assessment of the bridge safety status, possible waste of resources, and it may also not be able to fully utilize the advantages of preventive maintenance, increasing the risk of bridge emergencies. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a method and system for bridge surface crack detection based on image data processing, which can effectively solve the problems involved in the above background art.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: In the first aspect of the present invention, a method for detecting cracks on the surface of a bridge based on image data processing is provided, including the following steps: Analyze the basic information of the bridge to obtain the characteristic signal of the bridge's basic information. Based on the characteristic signal of the bridge's basic information, match the deployment scheme of the bridge detection camera; Based on the deployment scheme of the bridge detection camera, obtain the image quality data of each image captured by the bridge detection camera. Based on the image quality data of each image captured by the bridge detection camera, determine whether the image quality is qualified: If the image quality is unqualified, re-capture the image; If the image quality is qualified, extract each suspected crack area; Analyze each suspected crack area to determine whether the suspected crack area is a crack on the surface of the bridge: If the suspected crack area is not a crack on the surface of the bridge, mark the suspected crack area as an interference area; If the suspected crack area is a crack on the surface of the bridge, mark the suspected crack area as a crack area; Based on the judgment results of each suspected crack area, analyze the crack state on the surface of the bridge to determine the detection level result of the cracks on the surface of the bridge.

[0006] As a further method, analyze the basic information of the bridge to obtain the characteristic signal of the bridge's basic information. Based on the characteristic signal of the bridge's basic information, match the deployment scheme of the bridge detection camera. The specific analysis process is as follows: Obtain the basic information data of the bridge. The basic information data of the bridge specifically includes the total span of the bridge, the total length of the bridge deck, the total width of the bridge deck, and the total number of years since the bridge was put into use; Based on the obtained basic information data of the bridge, comprehensively analyze to obtain the characteristic signal of the bridge's basic information. The characteristic signal of the bridge's basic information is used as the analysis basis for matching the deployment scheme of the bridge detection camera; Store the characteristic signal of the bridge's basic information as a specified label, compare the specified label with each set label stored in the database to obtain the set label corresponding to the specified label, and obtain the deployment scheme of the bridge detection camera corresponding to the set label stored in the database.

[0007] As a further method, obtain the image quality data of each image captured by the bridge detection camera. Based on the image quality data of each image captured by the bridge detection camera, determine whether the image quality is qualified. The specific analysis process is as follows: Obtain the image quality data of each image captured by the bridge detection camera. The image quality data of each image captured by the bridge detection camera specifically includes image contrast, image signal-to-noise ratio, and the proportion of the occluded area of the image; Based on the obtained image quality data of each image captured by the bridge detection camera, comprehensively analyze to obtain the image quality analysis factor. The image quality analysis factor is used as the analysis basis for determining whether the image quality is qualified; Compare the image quality analysis factor with the image quality analysis threshold stored in the database; If the image quality analysis factor is lower than the image quality analysis threshold, the image quality corresponding to the image quality analysis factor is unqualified, and re-capture the image; If the image quality analysis factor is not lower than the image quality analysis threshold, the image quality corresponding to the image quality analysis factor is qualified, and extract each suspected crack area.

[0008] As a further method, the image quality analysis factor, and the specific analysis process is as follows:

[0009] ;

[0010] In the formula, is the image quality analysis factor, is the image contrast, is the image signal-to-noise ratio, is the proportion of the occluded area of the image, is the set compensation factor, is the set compensation factor, is the set compensation factor.

[0011] As a further method, extract each suspected crack area, and the specific analysis process is as follows: For each image with qualified quality, mark the gray value of the pixel with coordinates in the th image with qualified quality as , is the number of each image with qualified quality, , represents the number of images with qualified quality. For an image with a size of , ranges from 0 to , ranges from 0 to ; Use discrete difference to approximately calculate the approximate values of the partial derivatives in the and directions: In the direction, the partial derivative is approximately: - ; where is the approximate value of the partial derivative of the direction at the coordinates in the th image with qualified quality, is the gray value of the pixel with coordinates , is the gray value of the pixel with coordinates ; In the direction, the partial derivative is approximately: - ; where is the approximate value of the partial derivative of the direction at the coordinates in the th image with qualified quality, is the gray value of the pixel with coordinates The gray value of the pixel, is the gray value of the pixel with coordinates ; The gradient magnitude of the pixel with coordinates in the th qualified image is calculated using the Euclidean distance, ; The gradient direction of the pixel with coordinates in the th qualified image is ; When and , , when and , ; Compare the gradient magnitude of the pixel with coordinates in the th qualified image with the gradient magnitude threshold stored in the database, and compare the gradient direction of the pixel with coordinates in the th qualified image with the gradient direction defined range stored in the database; If the gradient magnitude of the pixel with coordinates in the th qualified image is not lower than the gradient magnitude threshold, and the gradient direction of the pixel with coordinates in the th qualified image belongs to the gradient direction defined range, then mark the pixel with coordinates in the th qualified image as a suspected crack pixel point; Otherwise, mark the pixel with coordinates in the th qualified image as a normal pixel point; Connect adjacent suspected crack pixel points into each suspected crack area.

[0012] As a further method, analyze each suspected crack area. The specific analysis process is as follows: Obtain the suspected crack area feature data. The suspected crack area feature data specifically includes the aspect ratio deviation rate of the suspected crack area, the perimeter deviation rate of the suspected crack area, and the circularity of the suspected crack area; Based on the obtained suspected crack area feature data, comprehensively analyze to obtain the suspected crack area feature factor. The suspected crack area feature factor is used as the analysis basis for judging whether the suspected crack area is a crack on the bridge surface;

[0013] The suspected crack area feature factor, the specific analysis process is as follows:

[0014] ;

[0015] In the formula, is the characteristic factor of the suspected crack area, is the aspect ratio deviation rate of the suspected crack area, is the perimeter deviation rate of the suspected crack area, is the circularity of the suspected crack area, is the set compensation factor of, is the set compensation factor of, is the set compensation factor of, where e is the natural constant.

[0016] As a further method, to determine whether the suspected crack area is a crack on the bridge surface, the specific analysis process is as follows: Compare the characteristic factor of the suspected crack area with the characteristic threshold of the suspected crack area stored in the database; if the characteristic factor of the suspected crack area is lower than the characteristic factor of the suspected crack area, then the suspected crack area corresponding to the characteristic factor of the suspected crack area is not a crack on the bridge surface, and mark this suspected crack area as an interference area; if the characteristic factor of the suspected crack area is not lower than the characteristic factor of the suspected crack area, then the suspected crack area corresponding to the characteristic factor of the suspected crack area is a crack on the bridge surface, and mark this suspected crack area as a crack area.

[0017] As a further method, based on the judgment results of each suspected crack area, analyze the crack state on the bridge surface to determine the crack detection level result of the bridge surface. The specific analysis process is as follows: Based on the judgment results of each suspected crack area, obtain the crack state data on the bridge surface. The crack state data on the bridge surface specifically includes the total number of cracks on the bridge surface, the average length of the cracks on the bridge surface, and the maximum density of the cracks on the bridge surface; based on the obtained crack state data on the bridge surface, comprehensively analyze to obtain the crack state signal on the bridge surface. The crack state signal on the bridge surface is used as the analysis basis for determining the crack detection level result of the bridge surface; compare the crack state signal on the bridge surface with the crack state threshold stored in the database; if the crack state signal on the bridge surface is not lower than the crack state threshold, then the crack detection level result corresponding to the crack state signal on the bridge surface is level two; if the crack state signal on the bridge surface is lower than the crack state threshold, then the crack detection level result corresponding to the crack state signal on the bridge surface is level one, and issue a danger warning for the crack state on the bridge surface.

[0018] As a further method, for the crack state signal on the bridge surface, the specific analysis process is as follows:

[0019] ;

[0020] In the formula, is the crack state signal on the bridge surface, is the total number of cracks on the bridge surface, is the average length of cracks on the bridge surface, is the maximum density of cracks on the bridge surface, is the set compensation factor of is the set compensation factor of is the set compensation factor of, where e is the natural constant.

[0021] The second aspect of the present invention provides a bridge surface crack detection system based on image data processing, including a camera deployment scheme matching module, an image quality qualification judgment module, a suspected crack area analysis module, and a crack detection level result determination module, where: The camera deployment scheme matching module is used to analyze the basic information of the bridge to obtain the basic information characteristic signal of the bridge, and based on the basic information characteristic signal of the bridge, match the bridge detection camera deployment scheme; The image quality qualification judgment module is used to obtain the image quality data of each image taken by the bridge detection camera based on the bridge detection camera deployment scheme, and based on the image quality data of each image taken by the bridge detection camera, judge whether the image quality is qualified: if the image quality is unqualified, re-acquire the image; if the image quality is qualified, extract each suspected crack area; The suspected crack area analysis module is used to analyze each suspected crack area to judge whether the suspected crack area is a crack on the bridge surface: if the suspected crack area is not a crack on the bridge surface, mark the suspected crack area as an interference area; if the suspected crack area is a crack on the bridge surface, mark the suspected crack area as a crack area; The crack detection level result determination module is used to analyze the bridge surface crack state based on the judgment results of each suspected crack area and determine the bridge surface crack detection level result.

[0022] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0023] (1) By providing a bridge surface crack detection method and system based on image data processing, the present invention matches the camera deployment scheme by analyzing the basic information of the bridge, and arranges the position and angle of the detection camera in a targeted manner. Avoid detection blind spots caused by unreasonable camera layout and improve the integrity of detection. The multi-dimensional image quality evaluation method can ensure that the images used for subsequent analysis have sufficient quality. Precisely locate the areas where cracks may exist, and perform multi-faceted feature analysis on the suspected crack areas, including the aspect ratio deviation rate, perimeter deviation rate, circularity, etc. of the crack area, to judge whether it is a real crack on the bridge surface and reduce misjudgment. Analyze the bridge surface crack state based on the judgment results of each suspected crack area and determine the detection level result. This can provide an intuitive and quantitative situation of the bridge surface cracks.

[0024] (2) By extracting each suspected crack area, the present invention determines whether the suspected crack area is a crack on the bridge surface based on the aspect ratio deviation rate, perimeter deviation rate, and circularity of the suspected crack area, avoiding misidentifying some stains or surface textures similar to cracks as cracks, thereby effectively distinguishing real cracks from other interference factors and greatly improving the accuracy of judgment. Characteristics such as aspect ratio deviation rate, perimeter deviation rate, and circularity are closely related to the essential shape attributes of cracks. During multiple detections, these shape characteristics can be used as stable judgment bases for repeated verification, improving the credibility of judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The present invention will be further described with reference to the accompanying drawings. However, the embodiments in the drawings do not constitute any limitation to the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the following drawings.

[0026] Figure 1 It is a schematic flowchart of the method steps of the present invention.

[0027] Figure 2 It is a schematic diagram of the connection of system modules of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0029] Referring to Figure 1 As shown, a method for detecting cracks on the bridge surface based on image data processing according to the first aspect of the present invention includes: analyzing the basic information of the bridge to obtain the basic information characteristic signal of the bridge, and matching the bridge detection camera deployment plan based on the basic information characteristic signal of the bridge.

[0030] The specific analysis process is as follows: Obtain the basic information data of the bridge, and the basic information data of the bridge specifically includes the total span of the bridge, the total length of the bridge deck, the total width of the bridge deck, and the total number of years since the bridge was put into use; based on the obtained basic information data of the bridge, comprehensively analyze to obtain the basic information characteristic signal of the bridge, and the basic information characteristic signal of the bridge is used as the analysis basis for matching the bridge detection camera deployment plan; store the basic information characteristic signal of the bridge as a specified label, compare the specified label with each set label stored in the database to obtain the set label corresponding to the specified label, and obtain the bridge detection camera deployment plan corresponding to the set label stored in the database.

[0031] The total span of a bridge refers to the sum of the clear spans of each hole in a multi-hole bridge, and is also known as the bridge aperture. The clear span is the clear distance between two adjacent piers (or abutments) at the design flood level. The total length of the bridge deck refers to the distance between the front ends of the side walls or flared walls of one abutment and those of the other abutment at the two ends of the bridge, that is, the length of the top surface of the carriageway (including the sidewalk part) of the superstructure of the bridge. The total width of the bridge deck refers to the total lateral width of the part of the bridge available for vehicle and pedestrian passage, including the carriageway width, the width of the median strip, the sidewalk width, etc. The total number of years since the bridge was put into use is the number of complete years elapsed from the date when the bridge was officially completed, opened to traffic and put into operation to the current date.

[0032] By analyzing the characteristic signals of the basic information of the bridge and matching the corresponding deployment plan, it can be ensured that the camera is placed at the most critical position where it can effectively capture the conditions of various parts of the bridge. For large bridges with a long total span and a wide bridge deck, cameras can be arranged at appropriate intervals and angles to comprehensively cover the areas to be detected, avoiding detection blind spots, so that more valuable bridge status data can be quickly and accurately obtained during a single detection process, reducing the situation of repeatedly adjusting the camera position or conducting multiple detections, and improving the efficiency of the overall detection work. Different basic information of the bridge determines different resource configurations such as the number and type of cameras required. According to the matched deployment plan, it can be accurately determined whether a high-definition long-focus lens camera is needed for long-distance observation, or a wide-angle camera is needed to cover a wider bridge deck area, etc., as well as the specific quantity, avoiding waste of resources and enabling the detection work to be carried out efficiently and orderly. Each bridge has unique structural characteristics and parts prone to problems due to its different basic information. For old bridges that have been in use for many years, some key connection parts may be more prone to aging, diseases, etc. The corresponding deployment plan will focus on monitoring these key parts, and through reasonable camera arrangement, clear and continuous observation of these key areas can be achieved, which helps to more accurately detect potential hidden dangers such as minor damages and diseases in the bridge, ensuring that the detection results can truly reflect the actual condition of the bridge and improving the detection quality.

[0033] The characteristic signals of the basic information of the bridge, the specific analysis process is as follows:

[0034] ;

[0035] In the formula, is the characteristic signal of the basic information of the bridge, is the total span of the bridge, is the total length of the bridge deck, is the total width of the bridge deck, is the total number of years since the bridge was put into use, is the set compensation factor, is the set compensation factor, is a set compensation factor of is a set compensation factor of, where e is the natural constant.

[0036] It should be noted that the above-mentioned basic information characteristic signal of the bridge is calculated through the total span of the bridge, the total length of the bridge deck, the total width of the bridge deck, and the total number of years since the bridge was put into use. For , , , , normalization processing is carried out. The four factors of the total span of the bridge, the total length of the bridge deck, the total width of the bridge deck, and the total number of years since it was put into use describe the basic situation of the bridge from two main dimensions: the size of the bridge and the service time. The basic information characteristic signal of the bridge can be used as an important basis for deploying bridge detection cameras. For bridges of different scales and usage conditions, the key detection parts and areas are different. This matching method based on characteristic signals can adapt to various types of bridges. Whether it is a beam bridge, an arch bridge or a cable-stayed bridge and other bridges with different structural forms, as long as the characteristic signal is calculated through its basic information, the corresponding camera deployment plan can be found in the database or a reasonable deployment strategy can be formulated for new types of bridges, ensuring that the arrangement of the detection cameras can fully consider the specific situation of the bridge.

[0037] It should be noted that the above-mentioned set , , , compensation factors are obtained from the database. According to the historical data, a mapping set of the total span of the bridge, the total length of the bridge deck, the total width of the bridge deck, the total number of years since the bridge was put into use and the , , , compensation factors is obtained to get the current , , , corresponding , q , , compensation factors.

[0038] It should be noted that in the following text , , , , , , , , They are also obtained through the mapping set of historical data and compensation factors established in the database, that is, the corresponding compensation factor is obtained according to the current data.

[0039] Based on the bridge inspection camera deployment plan, obtain the image quality data of each image taken by the bridge inspection camera. Based on the image quality data of each image taken by the bridge inspection camera, determine whether the image quality is qualified: if the image quality is unqualified, re-acquire the image; if the image quality is qualified, extract each suspected crack area.

[0040] The specific analysis process is as follows: obtain the image quality data of each image taken by the bridge inspection camera. The image quality data of each image taken by the bridge inspection camera specifically includes image contrast, image signal-to-noise ratio, and the proportion of the occluded area of the image. Based on the obtained image quality data of each image taken by the bridge inspection camera, comprehensively analyze to obtain an image quality analysis factor, and the image quality analysis factor is used as an analysis basis for judging whether the image quality is qualified. Compare the image quality analysis factor with the image quality analysis threshold stored in the database. If the image quality analysis factor is lower than the image quality analysis threshold, the image quality corresponding to the image quality analysis factor is unqualified, and re-acquire the image; if the image quality analysis factor is not lower than the image quality analysis threshold, the image quality corresponding to the image quality analysis factor is qualified, and extract each suspected crack area.

[0041] Image contrast refers to the degree of difference between the brightest and darkest areas in an image. It reflects the distribution of colors or gray values in the image and is used to measure the distinguishability of different parts of the image. Image signal-to-noise ratio is the ratio of signal to noise. In an image, the signal refers to the real and useful image information (such as the shape and color of the bridge structure), and the noise refers to the interference information generated due to various factors (such as the thermal noise of the camera sensor, uneven distribution of light, electromagnetic interference, etc.). The proportion of the occluded area of the image is the ratio of the area of the bridge part occluded by other objects in the image to the area that the entire bridge should display in the image.

[0042] For the image quality analysis factor, the specific analysis process is as follows:

[0043] ;

[0044] In the formula, is the image quality analysis factor, is the image contrast, is the image signal-to-noise ratio, is the proportion of the occluded area of the image, is the set compensation factor, is the set compensation factor, is the set compensation factor.

[0045] It should be noted that the above image quality analysis factors are calculated through image contrast, image signal-to-noise ratio, and the proportion of the occluded area of the image. For , , normalization processing is carried out. Sufficient contrast can clearly display the details of the bridge structure (such as tiny cracks, texture changes of materials, etc.). A high signal-to-noise ratio means that the useful information in the image (such as real structural features) can be better separated from the background noise (such as light interference, noise of the camera sensor itself, etc.), reducing misjudgment cases caused by poor image quality and providing an accurate data basis for subsequent bridge detection. Only images with qualified quality can ensure the effectiveness of the extraction of suspected crack areas. If the image is blurred, has low contrast, or is severely occluded, real cracks may be missed or false positive suspected crack areas may be generated. On the basis of high-quality images, it is possible to more accurately extract suspected crack areas, which is crucial for timely discovering potential safety hazards in the bridge structure. Judging whether the quality is qualified in a timely manner after image acquisition can avoid wasting time on subsequent analysis of unqualified images. The safety assessment of the bridge largely depends on the reliability of the detection data. Qualified image quality ensures that the data such as the extracted suspected crack areas can truly reflect the actual condition of the bridge.

[0046] For each image with qualified quality, mark the gray value of the pixel with coordinates in the th image with qualified quality as , where is the coordinate of the pixel in the image, is the number of each image with qualified quality, , represents the number of images with qualified quality, represents the horizontal direction of the image, represents the vertical direction of the image. For an image with a size of , the value range of is from 0 to , and the value range of is from 0 to ; Use discrete difference to approximately calculate the approximate values of the partial derivatives in the and directions: In the direction, the partial derivative approximation is: - ; where is the approximate value of the partial derivative of the direction at the coordinate in the th image with qualified quality, and is the approximate value of the partial derivative at the coordinate The gray value of the pixel, is the gray value of the pixel with coordinates ; In the direction, the partial derivative approximation is: - ; where is the th qualified image coordinate in the direction is the partial derivative approximation value, is the gray value of the pixel with coordinates , is the gray value of the pixel with coordinates ; The gradient magnitude of the pixel with coordinates in the th qualified image can be calculated using the Euclidean distance, ; The gradient direction of the pixel with coordinates in the th qualified image can be obtained through ; When and , , when and , ; Compare the gradient magnitude of the pixel with coordinates in the th qualified image with the gradient magnitude threshold stored in the database, and compare the gradient direction of the pixel with coordinates in the th qualified image with the gradient direction defined range stored in the database; If the gradient magnitude of the pixel with coordinates in the th qualified image is not lower than the gradient magnitude threshold, and the gradient direction of the pixel with coordinates in the th qualified image belongs to the gradient direction defined range, then mark the pixel with coordinates in the th qualified image as a suspected crack pixel point; Otherwise, mark the pixel with coordinates in the th qualified image as a normal pixel point; Connect adjacent suspected crack pixel points into each suspected crack region.

[0047] By calculating the approximate partial derivatives of image pixels in the x and y directions, the gradient magnitude, and the gradient direction, the image can be analyzed in detail at the pixel level. This method utilizes the characteristic that cracks in the image cause significant changes in pixel gray values, that is, the gradient magnitude of the pixels around the cracks is large, and the gradient direction has a certain pattern. By comparing with the thresholds and defined ranges in the database, suspected crack pixel points can be accurately identified. Connecting adjacent suspected crack pixel points forms a suspected crack region, which overcomes the possible misjudgment problem of individual pixel points. Because in actual images, due to interference from factors such as noise, there may be individual pixel points whose gradient magnitude and direction accidentally meet the conditions of suspected cracks, but connecting adjacent suspected pixels can make a judgment from the perspective of the region, which is more in line with the actual situation, thus more accurately locating the possible crack regions.

[0048] In the calculation process, instead of simply judging cracks based on the gray value change of a single pixel, the gray value changes around the pixel are comprehensively considered (through partial derivative and gradient calculations). This can effectively filter out the random fluctuations of pixel gray values caused by factors such as noise. By comparing the gradient direction of the pixel with the defined range stored in the database, the characteristics of cracks in different directions are considered. Different types of bridge structures and damage causes may result in different direction distributions of cracks, and this method can better adapt to this diversity, and can also effectively detect cracks with irregular shapes and directions, further improving the detection accuracy.

[0049] Analyze each suspected crack region to determine whether the suspected crack region is a crack on the bridge surface: If the suspected crack region is not a crack on the bridge surface, mark the suspected crack region as an interference region; if the suspected crack region is a crack on the bridge surface, mark the suspected crack region as a crack region.

[0050] The specific analysis process is as follows: Obtain the characteristic data of the suspected crack region. The characteristic data of the suspected crack region specifically includes the aspect ratio deviation rate of the suspected crack region, the perimeter deviation rate of the suspected crack region, and the circularity of the suspected crack region; Based on the obtained characteristic data of the suspected crack region, comprehensively analyze to obtain the characteristic factor of the suspected crack region, and the characteristic factor of the suspected crack region is used as the analysis basis for judging whether the suspected crack region is a crack on the bridge surface;

[0051] The characteristic factor of the suspected crack region, the specific analysis process is as follows:

[0052] ;

[0053] In the formula, is the characteristic factor of the suspected crack region, is the aspect ratio deviation rate of the suspected crack region, is the perimeter deviation rate of the suspected crack region, is the circularity of the suspected crack area, is the set compensation factor, is the set compensation factor, is the set compensation factor, and \(e\) is the natural constant.

[0054] The aspect ratio deviation rate of the suspected crack area. The aspect ratio deviation rate refers to the degree of difference between the aspect ratio of the suspected crack area and the ideal crack model, and is expressed by the formula: aspect ratio deviation rate = |(actual aspect ratio - standard aspect ratio) / standard aspect ratio|×100%. The perimeter deviation rate of the suspected crack area. The perimeter deviation rate refers to the degree of deviation between the actual perimeter of the suspected crack area and the ideal perimeter calculated based on its shape, and the calculation formula is: perimeter deviation rate = |(actual perimeter - ideal perimeter) / ideal perimeter|×100%. The circularity of the suspected crack area is an index used to measure the degree of closeness of the shape to a circle. For a planar region, there are multiple calculation formulas for circularity. A common one is: circularity = 4π×(area of the region) / (perimeter²). For the suspected crack area, substitute its area and perimeter into this formula to calculate the circularity. The circularity of a circle is 1, and the more irregular the shape, the closer the circularity is to 0.

[0055] It should be explained that the above-mentioned characteristic factors of the suspected crack area are calculated through the aspect ratio deviation rate of the suspected crack area, the perimeter deviation rate of the suspected crack area, and the circularity of the suspected crack area. Normalize , , , and comprehensively analyze to obtain the characteristic factors through these three characteristic data of the aspect ratio deviation rate, perimeter deviation rate, and circularity of the suspected crack area, and evaluate the suspected crack area from multiple dimensions. For example, real cracks on the surface of a bridge usually have certain shape characteristics, such as a relatively large aspect ratio (for slender cracks), the deviation of the perimeter from the theoretical shape conforms to the growth law of actual cracks, and the circularity is relatively low (because cracks are rarely circular). This multi-dimensional analysis can avoid misjudgments that may occur when relying solely on a single characteristic for judgment, and more accurately judge whether the suspected crack area is a real crack on the bridge surface. In actual images, there may be some interference areas similar to cracks, such as light and shadow changes, material surface textures, etc. Using these characteristic data can effectively distinguish these interference areas from real cracks.

[0056] Compare the characteristic factor of the suspected crack area with the characteristic threshold of the suspected crack area stored in the database; if the characteristic factor of the suspected crack area is lower than the characteristic factor of the suspected crack area, the suspected crack area corresponding to the characteristic factor of the suspected crack area is not a crack on the bridge surface, and mark this suspected crack area as an interference area; if the characteristic factor of the suspected crack area is not lower than the characteristic factor of the suspected crack area, the suspected crack area corresponding to the characteristic factor of the suspected crack area is a crack on the bridge surface, and mark this suspected crack area as a crack area.

[0057] By comparing the characteristic factor of the suspected crack area with the characteristic threshold in the database, it is possible to effectively distinguish the true cracks on the bridge surface from the interference areas. This judgment method based on quantitative indicators avoids the subjectivity and uncertainty that may occur when relying solely on manual visual inspection. Using the pre-set characteristic threshold as the judgment basis provides a standardized process for determining the suspected crack area. Different inspectors can make judgments according to the same standard when facing the same image data, reducing misjudgments caused by factors such as personal experience and observation angle. This helps to maintain high accuracy and consistency throughout the bridge inspection process. The entire comparison and marking process can be automated by writing a program. The computer can quickly calculate the characteristic factors of a large number of suspected crack areas, compare them with the characteristic thresholds, and then automatically complete the marking work. Compared with manually checking and judging each suspected crack area one by one, the automated detection process can process more data in a short time, improving the work efficiency of bridge inspection, especially suitable for comprehensive and frequent inspection tasks of large bridges.

[0058] Based on the judgment results of each suspected crack area, analyze the crack state on the bridge surface to determine the detection grade result of the crack on the bridge surface.

[0059] The specific analysis process is as follows: Based on the judgment results of each suspected crack area, obtain the crack state data on the bridge surface. The crack state data on the bridge surface specifically includes the total number of cracks on the bridge surface, the average length of the cracks on the bridge surface, and the maximum density of the cracks on the bridge surface; based on the obtained crack state data on the bridge surface, comprehensively analyze to obtain the crack state signal on the bridge surface. The crack state signal on the bridge surface is used as the analysis basis for determining the detection grade result of the crack on the bridge surface; compare the crack state signal on the bridge surface with the crack state threshold stored in the database; if the crack state signal on the bridge surface is not lower than the crack state threshold, the detection grade result of the crack on the bridge surface corresponding to the crack state signal on the bridge surface is grade two; if the crack state signal on the bridge surface is lower than the crack state threshold, the detection grade result of the crack on the bridge surface corresponding to the crack state signal on the bridge surface is grade one, and issue a danger warning for the crack state on the bridge surface.

[0060] The total number of cracks on the bridge surface refers to the sum of the quantities of all cracks that meet the judgment criteria after inspecting the bridge surface. This quantity is obtained by counting the marked crack areas, and each independent area determined to be a crack on the bridge surface is counted as one crack. The average length of cracks on the bridge surface refers to the average value of the lengths of all cracks determined to be on the bridge surface. The calculation method is to add up the lengths of all cracks and then divide by the total number of cracks. The length measurement can be obtained by using image analysis software to measure the longest axis length or boundary length of each crack area according to the conversion relationship between pixels and actual dimensions. The maximum density of cracks on the bridge surface refers to the maximum value of the crack length per unit area in a certain local area of the bridge surface. Usually, based on the regional division scheme stored in the database, the bridge surface is divided into several small regional units, and then the crack density in each regional unit is calculated respectively (generally expressed by dividing the sum of the lengths of all cracks in the area by the area of the region), and finally the maximum value among them is taken as the maximum density of cracks on the bridge surface.

[0061] By considering key data such as the total number of cracks on the bridge surface, average length, and maximum density to generate a status signal, the situation of cracks on the bridge surface can be comprehensively evaluated from multiple perspectives. The total number of cracks reflects the distribution range of cracks, the average length reflects the overall scale of cracks, and the maximum density can indicate the degree of aggregation of cracks in certain local areas. This comprehensive analysis method avoids the one-sidedness that may occur when relying solely on a single piece of data for evaluation, making the evaluation result more accurately reflect the true situation of cracks on the bridge surface. Comparing the status signal of cracks on the bridge surface with the threshold value in the database can clearly divide the detection level of cracks on the bridge surface. This provides a clear basis for subsequent maintenance decisions, enabling maintenance personnel to take different measures according to the severity of the cracks. The secondary detection level may mean that more urgent repairs or further detailed inspections are required, while the primary detection level may only require regular monitoring. A timely warning system is crucial for ensuring the safe operation of the bridge, especially for those bridges with heavy traffic and important transportation tasks.

[0062] The status signal of cracks on the bridge surface, the specific analysis process is as follows:

[0063] ;

[0064] In the formula, is the status signal of cracks on the bridge surface, is the total number of cracks on the bridge surface, is the average length of cracks on the bridge surface, is the maximum density of cracks on the bridge surface, is the set compensation factor, is a set compensation factor, is a set compensation factor, where e is the natural constant.

[0065] It should be noted that the above-mentioned bridge surface crack state signal is calculated through the total number of bridge surface cracks, the average length of bridge surface cracks, and the maximum density of bridge surface cracks. For , , normalization processing is performed. The total number of bridge surface cracks, the average length, and the maximum density respectively reflect the situation of bridge surface cracks from different perspectives. The total number reflects the distribution breadth of the cracks, the average length reflects the scale of the cracks, and the maximum density highlights the severity of the cracks in the local area. These factors can be comprehensively considered to avoid one-sidedness caused by relying only on a single index for evaluation, so as to more comprehensively and accurately grasp the overall situation of bridge surface cracks. The bridge surface crack condition can be represented by a quantified value. In this way, in different bridge detection scenarios, whether it is the detection of the same bridge at different times or the comparison between different bridges, the evaluation can be based on this quantified state signal.

[0066] Referring to Figure 2 shown, the second aspect of the present invention provides a bridge surface crack detection system based on image data processing, including a camera deployment scheme matching module, an image quality qualification judgment module, a suspected crack area analysis module, and a crack detection level result determination module.

[0067] The camera deployment scheme matching module is used to analyze the basic information of the bridge to obtain the basic information characteristic signal of the bridge, and based on the basic information characteristic signal of the bridge, match the bridge detection camera deployment scheme.

[0068] The image quality qualification judgment module is used to obtain the image quality data of each image taken by the bridge detection camera based on the bridge detection camera deployment scheme, and based on the image quality data of each image taken by the bridge detection camera, judge whether the image quality is qualified: if the image quality is unqualified, re-acquire the image; if the image quality is qualified, extract each suspected crack area.

[0069] The suspected crack area analysis module is used to analyze each suspected crack area to judge whether the suspected crack area is a bridge surface crack: if the suspected crack area is not a bridge surface crack, mark the suspected crack area as an interference area; if the suspected crack area is a bridge surface crack, mark the suspected crack area as a crack area.

[0070] The crack detection level result determination module is used to analyze the bridge surface crack state based on the judgment results of each suspected crack area and determine the bridge surface crack detection level result.

[0071] The above content is only an example and illustration of the structure of the present invention. Those skilled in the art of this technology can make various modifications, supplements, or use similar methods for substitution to the specific embodiments described. As long as they do not deviate from the structure of the invention or exceed the scope defined by this claim book, they shall fall within the protection scope of the present invention.

Claims

1. A bridge surface crack detection method based on image data processing, characterized in that: The following steps are involved: Analyze the basic information of the bridge to obtain the characteristic signal of the basic information of the bridge, and match the deployment plan of the bridge detection camera based on the characteristic signal of the basic information of the bridge; Based on the deployment plan of bridge inspection cameras, obtain the quality data of each image taken by the bridge inspection camera, and judge whether the image quality is qualified based on the quality data of each image taken by the bridge inspection camera: If the image quality is unsatisfactory, re-acquire the image; If the image quality is qualified, each suspected crack area is extracted; Analyze each suspected crack area to determine whether the suspected crack area is a crack on the bridge surface: If the suspected crack area is not a crack on the bridge surface, the suspected crack area is marked as an interference area; If the suspected crack area is a crack on the bridge surface, the suspected crack area is marked as a crack area; Based on the judgment results of each suspected crack area, the crack status of the bridge surface is analyzed to determine the bridge surface crack detection grade result.

2. The bridge surface crack detection method based on image data processing according to claim 1 is characterized in that: The basic information of the bridge is analyzed to obtain the characteristic signal of the basic information of the bridge. Based on the characteristic signal of the basic information of the bridge, the deployment plan of the bridge detection camera is matched. The specific analysis process is as follows: Obtaining basic information data of the bridge, which specifically includes the total span of the bridge, the total length of the bridge deck, the total width of the bridge deck, and the total number of years the bridge has been in service; Based on the acquired bridge basic information data, a comprehensive analysis is performed to obtain the bridge basic information characteristic signal, which is used as the analysis basis for matching the bridge inspection camera deployment plan; The characteristic signal of the basic information of the bridge is stored as a designated tag, and the designated tag is compared with each set tag stored in the database to obtain the set tag corresponding to the designated tag, and the bridge detection camera deployment plan corresponding to the set tag stored in the database is obtained.

3. The bridge surface crack detection method based on image data processing according to claim 1 is characterized in that: The acquisition of image quality data of each image taken by the bridge inspection camera and the determination of whether the image quality is qualified based on the image quality data of each image taken by the bridge inspection camera are specifically analyzed as follows: Obtaining quality data of each image taken by the bridge inspection camera, wherein the quality data of each image taken by the bridge inspection camera specifically includes image contrast, image signal-to-noise ratio, and image blocked area ratio; Based on the acquired image quality data captured by the bridge inspection camera, the image quality analysis factor is obtained through comprehensive analysis. The image quality analysis factor is used as the analysis basis for judging whether the image quality is qualified. comparing the image quality analysis factor with an image quality analysis threshold stored in a database; If the image quality analysis factor is lower than the image quality analysis threshold, the image quality corresponding to the image quality analysis factor is unqualified, and the image is re-collected; If the image quality analysis factor is not lower than the image quality analysis threshold, the image quality corresponding to the image quality analysis factor is qualified, and each suspected crack area is extracted.

4. The bridge surface crack detection method based on image data processing according to claim 3 is characterized in that: The specific analysis process of the image quality analysis factor is as follows: ； In the formula, is the image quality analysis factor, is the image contrast, is the image signal-to-noise ratio, is the ratio of the image blocked area, For setting The compensation factor, For setting The compensation factor, For setting compensation factor.

5. The bridge surface crack detection method based on image data processing according to claim 1 is characterized in that: The specific analysis process of extracting each suspected crack area is as follows: For each qualified image, The coordinates of the image with good quality are The gray value of the pixel is marked as , is the number of each image with qualified quality, , Indicates the number of images of acceptable quality. For an image of size images, The value range is 0 to , The value range is 0 to ; Using discrete difference approximation and Approximate partial derivatives in direction: exist In the direction, the partial derivative is approximately: - ; in, for Direction The coordinates of the image with good quality are Partial derivative approximation, The coordinates are The gray value of the pixel, The coordinates are The gray value of the pixel; exist In the direction, the partial derivative is approximately: - ; in, for Direction The coordinates of the image with good quality are Partial derivative approximation, The coordinates are The gray value of the pixel, The coordinates are The gray value of the pixel; No. The coordinates of the image with good quality are The gradient magnitude of the pixel Using Euclidean distance to calculate, ; No. The coordinates of the image with good quality are The gradient direction of the pixel for ; when and hour, ,when and hour, ; The first The coordinates of the image with good quality are The gradient magnitude of the pixel is compared with the gradient magnitude threshold stored in the database, and the The coordinates of the image with good quality are Compare the gradient direction of the pixel with the gradient direction definition range stored in the database; Jordi The coordinates of the image with good quality are The gradient magnitude of the pixel is not less than the gradient magnitude threshold, and the The coordinates of the image with good quality are If the gradient direction of the pixel belongs to the gradient direction definition range, then The coordinates of the image with good quality are The pixels are marked as suspected crack pixels; Otherwise, the The coordinates of the image with good quality are The pixels are marked as regular pixels; Adjacent suspected crack pixels are connected to form suspected crack areas.

6. The bridge surface crack detection method based on image data processing according to claim 1 is characterized in that: The analysis of each suspected crack area is carried out, and the specific analysis process is as follows: Acquiring characteristic data of the suspected crack region, wherein the characteristic data of the suspected crack region specifically includes a deviation rate of the aspect ratio of the suspected crack region, a deviation rate of the perimeter of the suspected crack region, and a circularity of the suspected crack region; Based on the acquired characteristic data of the suspected crack area, the characteristic factor of the suspected crack area is obtained through comprehensive analysis, and the characteristic factor of the suspected crack area is used as the analysis basis for judging whether the suspected crack area is a crack on the bridge surface; Characteristic factors of suspected crack areas, the specific analysis process is as follows: ； In the formula, is the characteristic factor of the suspected crack area, is the deviation rate of the aspect ratio of the suspected crack area, is the deviation rate of the perimeter of the suspected crack area, is the circularity of the suspected crack area, For setting The compensation factor, For setting The compensation factor, For setting The compensation factor of , e is a natural constant.

7. The method for detecting cracks on a bridge surface based on image data processing according to claim 6, characterized in that: The specific analysis process of determining whether the suspected crack area is a bridge surface crack is as follows: comparing the characteristic factor of the suspected crack region with the characteristic threshold of the suspected crack region stored in the database; If the characteristic factor of the suspected crack region is lower than the characteristic factor of the suspected crack region, the suspected crack region corresponding to the characteristic factor of the suspected crack region is not a crack on the bridge surface, and the suspected crack region is marked as an interference region; If the characteristic factor of the suspected crack area is not lower than the characteristic factor of the suspected crack area, the suspected crack area corresponding to the characteristic factor of the suspected crack area is a crack on the bridge surface, and the suspected crack area is marked as a crack area.

8. The bridge surface crack detection method based on image data processing according to claim 1 is characterized by: Based on the judgment results of each suspected crack area, the crack state of the bridge surface is analyzed to determine the bridge surface crack detection level result. The specific analysis process is as follows: Based on the judgment results of each suspected crack area, the bridge surface crack status data is obtained, and the bridge surface crack status data specifically includes the total number of bridge surface cracks, the average length of the bridge surface cracks, and the maximum density of the bridge surface cracks; Based on the acquired bridge surface crack state data, a bridge surface crack state signal is obtained through comprehensive analysis, and the bridge surface crack state signal is used as an analysis basis for determining the bridge surface crack detection grade result; comparing the bridge surface crack state signal with the bridge surface crack state threshold stored in the database; If the bridge surface crack state signal is not lower than the bridge surface crack state threshold, the bridge surface crack detection level result corresponding to the bridge surface crack state signal is level 2; If the bridge surface crack state signal is lower than the bridge surface crack state threshold, the bridge surface crack detection level result corresponding to the bridge surface crack state signal is level one, and a bridge surface crack state danger warning is issued.

9. The bridge surface crack detection method based on image data processing according to claim 8 is characterized in that: The specific analysis process of the bridge surface crack state signal is as follows: ； In the formula, is the bridge surface crack status signal, is the total number of cracks on the bridge surface, is the average length of cracks on the bridge surface, is the maximum density of cracks on the bridge surface, For setting The compensation factor, For setting The compensation factor, For setting The compensation factor of , e is a natural constant.

10. A bridge surface crack detection system based on image data processing, applied to a bridge surface crack detection method based on image data processing according to any one of claims 1 to 9, characterized in that: It includes a camera deployment scheme matching module, an image quality qualification judgment module, a suspected crack area analysis module and a crack detection level result determination module, among which: The camera deployment scheme matching module is used to analyze the basic information of the bridge, obtain the characteristic signal of the basic information of the bridge, and match the bridge detection camera deployment scheme based on the characteristic signal of the basic information of the bridge; The image quality qualification judgment module is used to obtain the quality data of each image taken by the bridge detection camera based on the bridge detection camera deployment plan, and judge whether the image quality is qualified based on the quality data of each image taken by the bridge detection camera: If the image quality is unsatisfactory, re-acquire the image; If the image quality is qualified, each suspected crack area is extracted; The suspected crack area analysis module is used to analyze each suspected crack area to determine whether the suspected crack area is a bridge surface crack: If the suspected crack area is not a crack on the bridge surface, the suspected crack area is marked as an interference area; If the suspected crack area is a crack on the bridge surface, the suspected crack area is marked as a crack area; The crack detection level result determination module is used to analyze the crack status of the bridge surface based on the judgment results of each suspected crack area and determine the bridge surface crack detection level result.

Images