A bridge pier defect rapid detection method and system for a pier climbing robot

Abstract

The application provides a bridge pier defect rapid detection method and system for a pier climbing robot, and relates to the technical field of bridge detection, and comprises the following steps: obtaining the hardness and carbonization depth of a pier surface to determine the concrete strength grade and aggregate type, and adjusting the brightness and contrast of a camera; obtaining the roughness arithmetic mean deviation and flatness data, and adjusting the shooting parameters and camera position; dividing the top, middle and bottom regions to detect defects; and determining the grade based on the defect characteristics and generating a result. The application realizes rapid and accurate detection of bridge pier defects by dynamically adjusting the camera parameters and adaptively detecting in different regions, such as edge enhancement in the top region, grid comparison in the middle region and contour enhancement in the bottom region, and provides reliable data support for the safety evaluation of pier structures.

Description

Technical Field

[0001] This invention relates to the field of bridge inspection technology, and more specifically, to a rapid detection method and system for bridge pier defects using a pier-climbing robot. Background Technology

[0002] As bridge construction continues to expand, bridge piers, as key load-bearing structures, are exposed to the outdoor environment for extended periods, making them susceptible to erosion from wind and rain, water flow, and vehicle loads, leading to defects. Traditional manual inspection requires scaffolding or the use of climbing equipment, which is inefficient and poses safety risks. While pier-climbing robots are increasingly being used for pier inspection, existing methods are not adapted to the specific defect characteristics of different areas of the pier, nor do they dynamically adjust inspection parameters based on the surface properties of the pier, making it difficult to balance inspection efficiency and accuracy.

[0003] Traditional methods for detecting defects in bridge piers have several shortcomings: First, the detection parameters are fixed and do not dynamically adjust the camera parameters based on the concrete strength grade, aggregate type, surface roughness, and flatness of the pier, resulting in poor image quality and affecting the accuracy of defect identification. Second, the piers are not divided into regions for detection; a uniform detection method is used for weathering cracks at the top, honeycomb pitting in the middle, and spalling defects at the bottom, which is not targeted enough. Third, the determination of defect levels lacks a systematic standard and relies solely on human experience, which is highly subjective and makes it difficult to form standardized and reliable detection results.

[0004] Therefore, it is necessary to design a rapid detection method and system for bridge pier defects using a pier-climbing robot to solve the problems of low detection efficiency, poor accuracy, and non-standard results caused by existing technologies, such as detection parameters not being dynamically adapted to pier characteristics, detection methods not being adjusted according to specific regions, and the lack of unified standards for defect level determination. Summary of the Invention

[0005] In view of this, the present invention proposes a rapid detection method and system for bridge pier defects using a pier-climbing robot, aiming to solve the problems of low detection efficiency and poor accuracy caused by the lack of dynamic adjustment of camera parameters, lack of regional adaptation of detection methods, and lack of unified standards for defect level determination in existing technologies.

[0006] In one aspect, the present invention proposes a rapid detection method for bridge pier defects using a pier-climbing robot, comprising:

[0007] S1, obtain the surface hardness and carbonation depth of the pier column, obtain the concrete strength grade and aggregate type of the pier column based on the surface hardness and carbonation depth, and adjust the brightness and contrast of the camera mounted on the pier climbing robot based on the concrete strength grade and aggregate type.

[0008] S2, acquire the arithmetic mean deviation of roughness and flatness data of the pier surface, and adjust the shooting parameters of the camera based on the arithmetic mean deviation of roughness, and adjust the position of the camera based on the flatness data;

[0009] S3, divide the pier into inspection areas for defect detection, and adjust the defect detection method based on the inspection areas. The inspection areas include: top inspection area, middle inspection area and bottom inspection area;

[0010] S4. Determine the defect level of the bridge pier based on the defect characteristics of each detection area, and generate the bridge pier defect detection result.

[0011] Furthermore, when determining the concrete strength grade and aggregate type of the pier column based on the surface hardness and the carbonation depth, the following steps are included:

[0012] When the surface hardness is within a first range and the carbonation depth is within a first interval, the concrete strength grade is determined to be the first grade and the aggregate type is the first type.

[0013] When the surface hardness is within a first range and the carbonation depth is within a second range, the concrete strength grade is determined to be the second grade and the aggregate type is the second type.

[0014] When the surface hardness is in the first range and the carbonation depth is in the third range, the concrete strength grade is determined to be the third grade and the aggregate type is the third type.

[0015] When the surface hardness is in the second range and the carbonation depth is in the first range, the concrete strength grade is determined to be the fourth grade and the aggregate type is the fourth type.

[0016] When the surface hardness is within the second range and the carbonation depth is within the second interval, the concrete strength grade is determined to be the fifth grade and the aggregate type is the fifth type.

[0017] When the surface hardness is in the second range and the carbonation depth is in the third range, the concrete strength grade is determined to be grade six and the aggregate type is type six.

[0018] When the surface hardness is in the third range and the carbonation depth is in the first range, the concrete strength grade is determined to be grade seven and the aggregate type is type seven.

[0019] When the surface hardness is in the third range and the carbonation depth is in the second range, the concrete strength grade is determined to be the eighth grade and the aggregate type is the eighth type.

[0020] When the surface hardness is in the third range and the carbonation depth is in the third interval, the concrete strength grade is determined to be the ninth grade and the aggregate type is the ninth type.

[0021] Furthermore, when adjusting the brightness and contrast of the camera mounted on the climbing robot based on the concrete strength grade and the aggregate type, the following steps are included:

[0022] When the concrete strength grade is first grade and the aggregate type is first type, the brightness is adjusted to a first value and the contrast is adjusted to a first ratio.

[0023] When the concrete strength grade is the second grade and the aggregate type is the second type, the brightness is adjusted to the second value and the contrast is adjusted to the second ratio.

[0024] When the concrete strength grade is grade three and the aggregate type is type three, the brightness is adjusted to the third value and the contrast is adjusted to the third ratio.

[0025] When the concrete strength grade is fourth grade and the aggregate type is fourth type, the brightness is adjusted to the fourth value and the contrast is adjusted to the fourth ratio.

[0026] When the concrete strength grade is grade 5 and the aggregate type is type 5, the brightness is adjusted to the fifth value and the contrast is adjusted to the fifth ratio.

[0027] When the concrete strength grade is grade 6 and the aggregate type is type 6, the brightness is adjusted to the sixth value and the contrast is adjusted to the sixth ratio.

[0028] When the concrete strength grade is grade 7 and the aggregate type is type 7, the brightness is adjusted to the seventh value and the contrast is adjusted to the seventh ratio.

[0029] When the concrete strength grade is grade 8 and the aggregate type is type 8, the brightness is adjusted to the eighth value and the contrast is adjusted to the eighth ratio.

[0030] When the concrete strength grade is grade nine and the aggregate type is type nine, the brightness is adjusted to the ninth value and the contrast is adjusted to the ninth ratio.

[0031] Furthermore, when adjusting the camera's shooting parameters based on the roughness arithmetic mean deviation, the following steps are included:

[0032] When the roughness arithmetic mean deviation is less than or equal to the preset roughness judgment threshold, the camera is controlled to take pictures at a resolution of the first horizontal pixel number multiplied by the first vertical pixel number, and the texture filtering function is turned off.

[0033] When the roughness arithmetic mean deviation is greater than the preset roughness judgment threshold, the camera is controlled to take pictures at a resolution of the second horizontal pixel number multiplied by the second vertical pixel number, and the texture filtering function is enabled.

[0034] The number of second horizontal pixels is greater than the number of first horizontal pixels, and the number of second vertical pixels is greater than the number of first vertical pixels.

[0035] Furthermore, when adjusting the camera position based on the flatness data, the process includes:

[0036] When the flatness data shows that the height of the protrusion on the pier surface is greater than a preset protrusion height threshold and the depth of the depression on the pier surface is less than or equal to a preset depression depth threshold, the distance between the camera and the pier surface is increased by a first distance.

[0037] When the flatness data shows that the height of the protrusion on the pier surface is less than or equal to a preset protrusion height threshold and the depth of the depression on the pier surface is greater than a preset depression depth threshold, the distance between the camera and the pier surface is reduced by a second distance.

[0038] When the flatness data shows that the height of the protrusion on the pier surface is less than or equal to a preset protrusion height threshold and the depth of the depression on the pier surface is less than or equal to a preset depression depth threshold, the current position of the camera remains unchanged.

[0039] Furthermore, when dividing the pier into inspection areas for defect detection, and adjusting the defect detection method based on the inspection areas, the process includes:

[0040] Regarding the top detection area:

[0041] The edge enhancement mode of the camera on the climbing robot is activated, and the top detection area image is captured based on the shooting parameters and camera position adjusted in step S2 and using the edge enhancement mode.

[0042] Obtain the edge grayscale change rate of each edge in the top detection region image;

[0043] When the edge grayscale change rate is greater than the preset edge grayscale change rate threshold, it is determined that there are weathering cracks in the top detection area;

[0044] When the edge grayscale change rate is less than or equal to the preset edge grayscale change rate threshold, it is determined that there are no weathering cracks in the top detection area.

[0045] Furthermore, when dividing the pier into inspection areas for defect detection and adjusting the defect detection method based on the inspection areas, the method also includes:

[0046] Regarding the central detection area:

[0047] Activate the detail enhancement mode of the camera mounted on the climbing robot, and capture an image of the central detection area based on the shooting parameters and camera position adjusted in step S2 and using the detail enhancement mode.

[0048] The image of the central detection region is divided into several grid units using a grid partitioning comparison method;

[0049] When the gray value of a pixel in a certain grid cell is less than the first threshold and the proportion of the number of pixels with a gray value less than the first threshold to the total number of pixels in the current grid cell is greater than the first proportion, it is determined that the current grid cell has a honeycomb surface and the current grid is recorded as the first grid cell.

[0050] When the number of pixels with gray values ​​greater than or equal to the first threshold or the number of pixels with gray values ​​less than the first threshold in a certain grid cell is less than or equal to the first proportion of the total number of pixels in the current grid cell, it is determined that the current grid cell does not have a honeycomb surface and the current grid is recorded as the second grid cell.

[0051] When the proportion of the number of the first grid unit to the total number of grid units in the central detection area is greater than a preset regional defect proportion threshold and the proportion of the number of the second grid unit to the total number of grid units in the central detection area is less than or equal to the preset regional defect proportion threshold, it is determined that there is honeycomb pitting in the central detection area.

[0052] When the proportion of the number of the second grid cells to the total number of grid cells in the central detection area is greater than a preset regional defect proportion threshold and the proportion of the number of the first grid cells to the total number of grid cells in the central detection area is less than or equal to a preset regional defect proportion threshold, it is determined that there is no honeycomb surface in the central detection area.

[0053] Furthermore, when dividing the pier into inspection areas for defect detection and adjusting the defect detection method based on the inspection areas, the method also includes:

[0054] Regarding the bottom detection area:

[0055] The contour enhancement mode of the camera on the climbing robot is activated, and the bottom detection area image is captured based on the shooting parameters and camera position adjusted in step S2 and the contour enhancement mode is used.

[0056] The percentage of the area covered by the surface deposits in the bottom detection area relative to the total area of ​​the bottom detection area;

[0057] When the coverage area ratio is greater than the second ratio, the surface attachments of the bottom detection area are cleaned and the image of the bottom detection area is re-captured;

[0058] When the coverage area ratio is less than or equal to the second ratio, the defect features present in the bottom detection area are determined based on the area of ​​the continuous blank area.

[0059] When the area of ​​a continuous blank region in the bottom detection region image is greater than the first area threshold, it is determined that there is a peeling defect in the bottom detection region.

[0060] When the area of ​​a continuous blank region in the bottom detection region image is less than the first area threshold, it is determined that there is no peeling defect in the bottom detection region.

[0061] Furthermore, when determining the defect level of the bridge pier based on the defect characteristics of each of the detection areas and generating the bridge pier defect detection results, the process includes:

[0062] When the weathering cracks are present in the top detection area, the honeycomb surface is present in the middle detection area, and the spalling defects are present in the bottom detection area, the bridge pier defect level is determined to be Level 1.

[0063] When the weathering cracks are present in the top detection area, the honeycomb surface is present in the middle detection area, and the spalling defects are not present in the bottom detection area, the defect level of the bridge pier column is determined to be Level II.

[0064] When the weathering cracks are present in the top detection area, the honeycomb surface is not present in the middle detection area, and the peeling defects are present in the bottom detection area, the defect level of the bridge pier is determined to be level three.

[0065] When the top detection area does not have weathering cracks, the middle detection area has honeycomb pitting, and the bottom detection area has spalling defects, the bridge pier defect level is determined to be level four.

[0066] When the top detection area does not have weathering cracks, the middle detection area does not have honeycomb pitting, and the bottom detection area has peeling defects, the bridge pier defect level is determined to be level five.

[0067] When the top detection area does not have weathering cracks, the middle detection area has honeycomb pitting, and the bottom detection area does not have spalling defects, the bridge pier defect level is determined to be level six.

[0068] When the weathering cracks are present in the top detection area, the honeycomb surface is not present in the middle detection area, and the peeling defects are not present in the bottom detection area, the defect level of the bridge pier is determined to be level seven.

[0069] When the top detection area does not have weathering cracks, the middle detection area does not have honeycomb pitting, and the bottom detection area does not have spalling defects, the bridge pier defect level is determined to be level eight.

[0070] The defect levels and the defect features of each of the detection areas are integrated to generate a bridge pier defect detection result that includes the defect levels and the defect features of each of the detection areas.

[0071] Compared with existing technologies, the beneficial effects of this invention are as follows: The rapid defect detection method for bridge piers using a climbing robot determines the concrete strength grade and aggregate type based on the surface hardness and carbonation depth of the pier, dynamically adjusts camera brightness and contrast, and optimizes shooting parameters and positions by combining roughness and flatness to ensure image quality. Different detection methods are adapted for the top, middle, and bottom regions, such as enhancing the identification of weathering cracks at the top edge, detecting honeycomb pitting through grid comparison in the middle, and determining spalling defects by enhancing the bottom contour, thus improving the targeted nature of defect identification. Furthermore, the defect level is determined based on the defect characteristics of each region, and standardized results are generated, effectively solving the problems of low efficiency and poor accuracy in traditional detection methods, and providing reliable support for the safety assessment of bridge piers.

[0072] On the other hand, this invention proposes a rapid defect detection system for bridge piers using a pier-climbing robot, comprising:

[0073] The camera brightness adjustment module is used to obtain the surface hardness and carbonation depth of the pier column, obtain the concrete strength grade and aggregate type of the pier column based on the surface hardness and carbonation depth, and adjust the brightness and contrast of the camera mounted on the pier climbing robot based on the concrete strength grade and aggregate type.

[0074] The camera position adjustment module is used to acquire the arithmetic mean deviation of the roughness and the flatness data of the pier surface, and adjust the shooting parameters of the camera based on the arithmetic mean deviation of the roughness, and adjust the position of the camera based on the flatness data.

[0075] The regional defect detection module is used to divide the pier into detection areas for defect detection, and adjust the defect detection method based on the detection areas. The detection areas include: a top detection area, a middle detection area, and a bottom detection area.

[0076] The defect result generation module is used to determine the defect level of the bridge pier based on the defect characteristics of each detection area, and generate the bridge pier defect detection results.

[0077] It is understandable that the above-mentioned method and system for rapid detection of bridge pier defects using a climbing robot have the same beneficial effects, and will not be elaborated further here. Attached Figure Description

[0078] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0079] Figure 1 This is a flowchart of a rapid defect detection method for bridge piers using a pier-climbing robot, provided in an embodiment of the present invention.

[0080] Figure 2 This is a functional block diagram of a rapid defect detection system for bridge piers using a climbing robot, provided in an embodiment of the present invention. Detailed Implementation

[0081] Exemplary embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present invention and to fully convey the scope of the invention to those skilled in the art. It should be noted that, without conflict, the embodiments and features described herein can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0082] Reference Figure 1 As shown in some embodiments of this application, a rapid detection method for bridge pier defects using a pier-climbing robot includes:

[0083] S1, obtain the surface hardness and carbonation depth of the pier column, obtain the concrete strength grade and aggregate type of the pier column based on the surface hardness and carbonation depth, and adjust the brightness and contrast of the camera mounted on the pier climbing robot based on the concrete strength grade and aggregate type.

[0084] S2, acquire the arithmetic mean deviation of roughness and flatness data of the pier surface, and adjust the shooting parameters of the camera based on the arithmetic mean deviation of roughness, and adjust the position of the camera based on the flatness data;

[0085] S3, divide the pier into inspection areas for defect detection, and adjust the defect detection method based on the inspection areas. The inspection areas include: top inspection area, middle inspection area and bottom inspection area;

[0086] S4. Determine the defect level of the bridge pier based on the defect characteristics of each detection area, and generate the bridge pier defect detection result.

[0087] Specifically, surface hardness refers to the ability of the concrete surface of the pier to resist external force indentation or scratching. It is obtained by a rebound hammer mounted on the pier climbing robot. Specifically, the rebound hammer is used to strike the surface of the pier, and the surface hardness is determined based on the value indicated by the pointer of the rebound hammer after the strike. Carbonation depth refers to the depth of the carbonation layer formed from the surface to the interior of the concrete due to the chemical reaction between carbon dioxide in the atmosphere and calcium hydroxide in the concrete surface of the pier. It is obtained by drilling a concrete core sample from the surface of the pier using a drilling device mounted on the pier climbing robot, and then spraying phenolphthalein reagent on the cross-section of the core sample using a reagent spraying device mounted on the robot. The vertical distance from the boundary line between the area where the reagent does not discolor (carbonized part) and the area where it discolors (uncarbonized part) to the surface of the core sample is measured.

[0088] Specifically, concrete strength grade is a classification based on the standard value of the compressive strength of a concrete cube (the compressive strength measured at 28 days of standard curing using a standard core sample with a diameter of 10cm and a height-to-diameter ratio of 1:1), reflecting the concrete's load-bearing capacity. Aggregate type refers to the type of coarse aggregate that acts as the skeleton in concrete, mainly crushed stone (formed from crushed rock, with sharp edges and a rough surface) and pebbles (formed by natural water erosion, with a smooth surface and rounded shape), affecting the workability and durability of concrete.

[0089] Specifically, the roughness arithmetic mean deviation refers to the surface profile curve obtained by sliding a probe at a constant speed along the surface within a selected sampling length on the pier surface using a contact surface measuring device. This is done by recording the probe's displacement in the direction perpendicular to the surface in real time. Then, the profile centerline (the straight line within the sampling length that minimizes the sum of the squares of the distances from each point on the profile to this line) is determined. The absolute values ​​of the distances from all points on the profile to the profile centerline are calculated, and their arithmetic mean is taken as the roughness arithmetic mean deviation within that sampling length. The final value is obtained by averaging the results from multiple sampling lengths. The flatness data refers to the acquisition of three-dimensional coordinate data of a large number of measuring points by scanning the pier surface along multiple paths at preset intervals using a laser displacement sensor. A reference plane is formed by fitting a set of continuous measuring points (using the least squares method to minimize the sum of the squares of the distances from each point to the plane). The perpendicular distance between each measuring point and the reference plane is calculated, and the difference between the maximum and minimum values ​​of all distances is taken as the overall flatness data, or the arithmetic mean of the absolute values ​​of the distances is taken as the average flatness data.

[0090] Understandably, by determining the concrete strength grade and aggregate type based on the surface hardness and carbonation depth of the pier, dynamically adjusting the camera brightness and contrast, and optimizing the shooting parameters and position in conjunction with roughness and flatness, image quality is ensured. Different detection methods are adapted for the top, middle, and bottom areas, such as enhancing the identification of weathering cracks at the top edge, detecting honeycomb pitting through grid comparison in the middle, and determining spalling defects by enhancing the bottom contour, thus improving the targeting of defect identification. Furthermore, the defect level is determined based on the defect characteristics of each area, and standardized results are generated, effectively solving the problems of low efficiency and poor accuracy in traditional detection methods, and providing reliable support for the safety assessment of bridge piers.

[0091] In some embodiments of this application, when obtaining the concrete strength grade and aggregate type of the pier column based on the surface hardness and the carbonation depth, the following steps are included:

[0092] When the surface hardness is within a first range and the carbonation depth is within a first interval, the concrete strength grade is determined to be the first grade and the aggregate type is the first type.

[0093] When the surface hardness is within a first range and the carbonation depth is within a second range, the concrete strength grade is determined to be the second grade and the aggregate type is the second type.

[0094] When the surface hardness is in the first range and the carbonation depth is in the third range, the concrete strength grade is determined to be the third grade and the aggregate type is the third type.

[0095] When the surface hardness is in the second range and the carbonation depth is in the first range, the concrete strength grade is determined to be the fourth grade and the aggregate type is the fourth type.

[0096] When the surface hardness is within the second range and the carbonation depth is within the second interval, the concrete strength grade is determined to be the fifth grade and the aggregate type is the fifth type.

[0097] When the surface hardness is in the second range and the carbonation depth is in the third range, the concrete strength grade is determined to be grade six and the aggregate type is type six.

[0098] When the surface hardness is in the third range and the carbonation depth is in the first range, the concrete strength grade is determined to be grade seven and the aggregate type is type seven.

[0099] When the surface hardness is in the third range and the carbonation depth is in the second range, the concrete strength grade is determined to be the eighth grade and the aggregate type is the eighth type.

[0100] When the surface hardness is in the third range and the carbonation depth is in the third interval, the concrete strength grade is determined to be the ninth grade and the aggregate type is the ninth type.

[0101] Specifically, the first, second, and third ranges of surface hardness were obtained by using a rebound hammer mounted on a pier-climbing robot to conduct impact tests on a large number of different bridge pier sample surfaces. For each pier sample, multiple evenly distributed test points were selected on its surface. After each test point was impacted, the value indicated by the rebound hammer pointer was recorded. Outliers deviating from the normal range were removed, and valid data was retained. After collecting a massive amount of raw surface hardness data, combined with the actual distribution characteristics of pier surface hardness in engineering projects and testing requirements, these valid data were manually divided into three continuous and non-overlapping numerical ranges (e.g., the first range corresponds to rebound values ​​of 18-28, the second range corresponds to 28-38, and the third range corresponds to...). The range corresponds to 38-48, and the division principle is to cover most sample data and clearly distinguish different hardness levels. The first, second, and third intervals of carbonization depth are obtained by drilling standard core samples from the above-mentioned pier column samples with tested surface hardness using a pier-climbing robot. The robot is equipped with a drilling device to drill standard cylindrical core samples with a diameter of 10cm and a height-to-diameter ratio of 1:1 near the test point of each pier column sample. After cleaning the cross-section of the core sample, phenolphthalein reagent is evenly sprayed using a reagent spraying device. After the reagent color stabilizes (1-2 minutes), the vertical distance from the boundary line between the non-coloring area (carbonized part) and the coloring area (uncarbonized part) on the cross-section of the core sample to the surface of the core sample is measured. A large amount of carbonized material is collected. After obtaining the depth data, based on common conditions of carbonation degree in pier concrete and the requirements for testing and grading, these data were artificially divided into three continuous and non-overlapping depth intervals (e.g., the first interval 0-4mm, the second interval 4-8mm, and the third interval 8-12mm, ensuring that each interval reflects a typical degree of carbonation). The first to ninth grades of concrete strength were obtained by drilling standard core samples from the same pier sample used for testing surface hardness and carbonation depth: for each pier sample, a standard core sample with a diameter of 10cm and a height-to-diameter ratio of 1:1, originating from the same source as the carbonation depth test, was drilled and directly placed on a press for compressive strength testing (this size...). The compressive strength results of the core samples do not need to be converted and can be used directly. After collecting a large amount of core sample compressive strength data, and combining the engineering application scenarios and grading requirements of concrete bearing capacity, these strength data are artificially divided into nine levels (for example, the first level corresponds to a compressive strength of 14-19MPa, the second level to 19-24MPa, the third level to 24-29MPa, the fourth level to 29-34MPa, the fifth level to 34-39MPa, the sixth level to 39-44MPa, the seventh level to 44-49MPa, the eighth level to 49-54MPa, and the ninth level to 54-59MPa, with each level precisely matching the actual bearing capacity of the corresponding concrete).The aggregate types, from Type I to Type 9, are derived from the analysis of coarse aggregate in the aforementioned standard core samples of pier columns. Each core sample is crushed to separate the coarse aggregate that forms the framework (containing only crushed stone and pebbles). This is achieved by artificially refining the two main categories of crushed stone and pebbles, considering the material composition (e.g., granite, limestone), particle size range (e.g., 5-10mm, 10-20mm, 20-30mm), and surface characteristics (e.g., the sharpness of the edges of crushed stone, the smoothness of the pebbles). (For example, Type I). The classification is as follows: Type 1: 5-10mm granite crushed stone; Type 2: 10-20mm granite crushed stone; Type 3: 20-30mm granite crushed stone; Type 4: 5-10mm limestone crushed stone; Type 5: 10-20mm limestone crushed stone; Type 6: 20-30mm limestone crushed stone; Type 7: 5-10mm river pebbles; Type 8: 10-20mm river pebbles; Type 9: 20-30mm river pebbles. The classification ensures that each type reflects the different impacts of aggregates on the workability and durability of concrete.

[0102] Specifically, the correspondence between surface hardness (ranges 1-3) and carbonation depth (ranges 1-3) and concrete strength grades (grades 1-9) was obtained through extensive experiments: For each experimental pier sample, its surface hardness was first tested using a rebound hammer and assigned to the corresponding hardness range. Then, a standard core sample with a diameter of 10 cm and a height-to-diameter ratio of 1:1 was drilled from the same testing area of ​​the same pier. The carbonation depth of the core sample was measured and assigned to the corresponding carbonation depth range. The compressive strength of the core sample was also directly tested and assigned to the corresponding strength grade. By recording a large amount of correlation data between "same pier sample - surface hardness range - carbonation depth range - concrete strength grade", the stable concrete strength grades corresponding to different combinations of surface hardness ranges and carbonation depth ranges were observed. The strength grades of concrete were analyzed to establish a one-to-one correspondence. Similarly, the correspondence between surface hardness (ranges 1 to 3) and carbonation depth (ranges 1 to 3) and aggregate type (types 1 to 9) was obtained through extensive experiments. For each pier sample used in the experiment, its surface hardness was first tested and assigned to the corresponding range. Then, a standard core sample from the same area was drilled to measure the carbonation depth and assigned to the corresponding range. At the same time, the core sample was crushed to separate the coarse aggregate. After analyzing the aggregate material, particle size, and surface characteristics, it was assigned to the corresponding aggregate type. A large amount of correlation data between "same pier sample - surface hardness range - carbonation depth range - aggregate type" was recorded. The aggregate types that were stably corresponding to different combinations of surface hardness ranges and carbonation depth ranges were observed, and finally, a one-to-one correspondence was summarized.

[0103] Specifically, the determination of the concrete strength grade and aggregate type is made by looking up the correspondence between surface hardness and carbonation depth and the concrete strength grade, as well as the correspondence between surface hardness and carbonation depth and the aggregate type.

[0104] Understandably, by establishing a correlation between the basic characteristics of the pier and the detection adaptation parameters, and through a clear correspondence between ranges and intervals, the pier-climbing robot can accurately identify the material properties of the pier, providing accurate material information support for subsequent targeted adjustments to camera parameters and avoiding detection deviations caused by material misjudgment.

[0105] In some embodiments of this application, adjusting the brightness and contrast of the camera mounted on the climbing robot based on the concrete strength grade and the aggregate type includes:

[0106] When the concrete strength grade is first grade and the aggregate type is first type, the brightness is adjusted to a first value and the contrast is adjusted to a first ratio.

[0107] When the concrete strength grade is the second grade and the aggregate type is the second type, the brightness is adjusted to the second value and the contrast is adjusted to the second ratio.

[0108] When the concrete strength grade is grade three and the aggregate type is type three, the brightness is adjusted to the third value and the contrast is adjusted to the third ratio.

[0109] When the concrete strength grade is fourth grade and the aggregate type is fourth type, the brightness is adjusted to the fourth value and the contrast is adjusted to the fourth ratio.

[0110] When the concrete strength grade is grade 5 and the aggregate type is type 5, the brightness is adjusted to the fifth value and the contrast is adjusted to the fifth ratio.

[0111] When the concrete strength grade is grade 6 and the aggregate type is type 6, the brightness is adjusted to the sixth value and the contrast is adjusted to the sixth ratio.

[0112] When the concrete strength grade is grade 7 and the aggregate type is type 7, the brightness is adjusted to the seventh value and the contrast is adjusted to the seventh ratio.

[0113] When the concrete strength grade is grade 8 and the aggregate type is type 8, the brightness is adjusted to the eighth value and the contrast is adjusted to the eighth ratio.

[0114] When the concrete strength grade is grade nine and the aggregate type is type nine, the brightness is adjusted to the ninth value and the contrast is adjusted to the ninth ratio.

[0115] Specifically, camera brightness refers to the parameter controlling the brightness of the image when the camera captures it, while camera contrast refers to the parameter controlling the difference in brightness between bright and dark areas in the image. The first to ninth values ​​corresponding to camera brightness are specific brightness parameter values ​​matching the first to ninth combinations of concrete strength grade and aggregate type, respectively. The first to ninth ratios corresponding to camera contrast are specific contrast ratio parameters matching the above combinations, respectively. The specific method for obtaining these parameters is as follows: collect pier sample specimens covering different concrete strength grades (such as C15, C25, C35, etc.) and different aggregate types (such as river sand aggregate, crushed stone aggregate, slag aggregate, etc.). Under uniform environmental conditions, use the same camera mounted on the pier-climbing robot to take multiple sets of pictures of each sample. In each set of pictures, adjust the camera brightness knob (range set to 0-100 levels) and contrast adjustment key (range set to 0%-100%), and record the results simultaneously. Images are recorded at different brightness levels and contrast ratios. Professional inspectors then score the identifiability of typical defects (such as 0.2mm wide cracks and 5mm² holes) in the images (1-10 points). For each sample, the brightness level and contrast ratio corresponding to the highest defect identifiability score are selected as the matching parameters for that sample. Finally, through statistical analysis, samples with similar parameters are grouped into nine categories, and the first to ninth values ​​and ratios are determined accordingly. In practical applications, if the detected pier column concrete strength grade is C15 and the aggregate is river sand (first combination), the first value (such as level 25) of brightness and the first ratio (such as 55%) of contrast are used. The image captured at this time can clearly show the surface cracks. If it is C35 and the aggregate is slag (sixth combination), because the surface of slag aggregate is darker, the sixth value (such as level 40) of brightness and the sixth ratio (such as 65%) of contrast are used to avoid the missed detection of holes caused by the image being too dark.

[0116] Understandably, for different concrete strength grades and aggregate types, adjustment schemes for camera brightness and contrast were defined, achieving precise adaptation of detection parameters to pier material. By presetting corresponding parameters, it is ensured that the camera can capture clear images on the surface of piers of different materials, providing a high-quality imaging foundation for subsequent defect identification and improving the initial accuracy of defect identification.

[0117] In some embodiments of this application, adjusting the camera's shooting parameters based on the roughness arithmetic mean deviation includes:

[0118] When the roughness arithmetic mean deviation is less than or equal to the preset roughness judgment threshold, the camera is controlled to take pictures at a resolution of the first horizontal pixel number multiplied by the first vertical pixel number, and the texture filtering function is turned off.

[0119] When the roughness arithmetic mean deviation is greater than the preset roughness judgment threshold, the camera is controlled to take pictures at a resolution of the second horizontal pixel number multiplied by the second vertical pixel number, and the texture filtering function is enabled.

[0120] The number of second horizontal pixels is greater than the number of first horizontal pixels, and the number of second vertical pixels is greater than the number of first vertical pixels.

[0121] Specifically, the preset roughness threshold is a critical physical quantity used to distinguish between high and low surface roughness of the pier and thus determine the camera's adaptive shooting parameters. Its value is directly related to the clarity of defect recognition. It is necessary to ensure that when the surface roughness of the pier is below this threshold, a lower resolution can be used to clearly identify defects. When it is above this threshold, a higher resolution needs to be switched to avoid small defects being covered by surface texture. The first horizontal pixel count and the first vertical pixel count are the number of pixels contained in the horizontal and vertical directions of the image when the camera shoots at a normal resolution. The two together constitute the specific parameters of the normal resolution and must meet the recognition requirements of common defects (such as wide cracks and obvious peeling) when the surface roughness of the pier is low. The second horizontal pixel count and the second vertical pixel count are the number of pixels contained in the horizontal and vertical directions of the image when the camera shoots at a higher resolution. The second horizontal pixel count is greater than the first horizontal pixel count, and the second vertical pixel count is greater than the first vertical pixel count. The higher resolution constituted by the two must meet the recognition requirements of small defects (such as fine weathering cracks and small honeycomb pits) when the surface roughness of the pier is high, ensuring that such defects can still be clearly captured in complex surface textures.

[0122] Specifically, the preset roughness judgment threshold is obtained through the following specific operations: First, collect actual or simulated samples of piers with different roughness levels, covering typical surface states from smooth to rough; then, take pictures of each sample at multiple different resolutions (including conventional resolution and higher resolution to be determined later) to obtain surface images of each sample at different resolutions; next, perform identification tests on standard defects in the images (such as cracks of preset width and holes of preset area), and record the accuracy and clarity of defect identification at different resolutions for samples with different roughness; finally, analyze the data to find the critical roughness value—when the sample roughness is below this value, the defect identification accuracy at conventional resolution reaches the preset standard; when it is above this value, the accuracy at conventional resolution is below the standard while the accuracy at higher resolution meets the standard. This critical value is the preset roughness judgment threshold. The first horizontal pixel count and the first vertical pixel count are obtained through the following operations: First, the minimum recognition size requirements for common defects on the pier surface (such as cracks with a width ≥ 0.2 mm and peeling with an area ≥ 5 cm²) are determined; then, based on the camera hardware performance (such as maximum pixels and sensor size), multiple combinations of horizontal and vertical pixel counts are set as candidate parameters for conventional resolution; then, standard samples containing common defects are photographed using each candidate parameter, and the pixel ratio of defects in the image is measured after obtaining the image. The parameter combination that allows the pixel ratio of common defects to meet the minimum pixel requirement of the recognition algorithm (such as defects occupying at least 5 consecutive pixels in the image) and the image data volume to be moderate (avoiding excessive storage and processing resources) is selected and determined as the first horizontal pixel count and the first vertical pixel count. The second horizontal pixel count and the second vertical pixel count are obtained through a similar process: First, the minimum recognition size requirement for minute defects (such as slits with a width ≥ 0.1 mm or tiny honeycombs with an area ≥ 1 cm²) is defined; then, based on the upper limit of the camera hardware, multiple combinations of horizontal and vertical pixel counts higher than the conventional resolution are set as candidate parameters; then, standard samples containing minute defects are photographed using each candidate parameter, and the recognition accuracy of minute defects is tested and recorded; finally, the parameter combination that enables the recognition accuracy of minute defects to reach the preset standard and is within the allowable range of hardware performance (such as shooting speed and data transmission rate) is selected and determined as the second horizontal pixel count and the second vertical pixel count.

[0123] Specifically, taking the practical application of defect detection in concrete piers of highway bridges as an example, the preset parameters are obtained through the following specific operations: First, 10 concrete piers with different surface conditions are selected from existing bridges as samples. Among them, 3 have smooth surfaces (after grinding, with an arithmetic mean roughness deviation of approximately 0.3 μm), 4 have medium rough surfaces (naturally weathered without grinding, with a deviation of approximately 0.6-0.8 μm), and 3 have rough surfaces (poor surface treatment during construction, with a deviation of approximately 1.0-1.2 μm). Standard defects are artificially created on the surface of each pier (0.2 mm wide defects are created for smooth samples). For samples with cracks and 5cm² spalling, a 0.15mm wide crack and 3cm² honeycomb pattern were created; for samples with coarse texture, a 0.1mm wide crack and 1cm² honeycomb pattern were created. Then, using an industrial camera mounted on a climbing robot, each sample was photographed with five pixel combinations: 2000×1500, 2500×1800, 3000×2200, 3500×2600, and 4000×3000. Each parameter combination was photographed three times to obtain clear images. Next, standard defects in the images were manually labeled, and the defect identification accuracy under each parameter was statistically analyzed using detection software. It was found that when the arithmetic mean deviation of sample roughness was ≤0.6μm, the defect recognition accuracy of the 2000×1500 pixel combination was ≥95% (e.g., 100% recognition rate for a 0.2mm crack in a smooth sample). When the deviation was >0.6μm, the accuracy of this combination dropped to below 80% (e.g., only 75% recognition rate for a 0.1mm crack in a rough sample). However, the defect recognition accuracy of samples with a deviation >0.6μm in the 3500×2600 pixel combination was ≥92% (93% recognition rate for a 0.1mm crack in a rough sample). Finally, the preset roughness judgment threshold was determined to be 0.6μm. The first horizontal pixel count is 2000, the first vertical pixel count is 1500 (standard resolution), the second horizontal pixel count is 3500, and the second vertical pixel count is 2600 (higher resolution). This parameter combination was verified in actual bridge inspection to meet both the requirements for rapid identification of common defects (approximately 15 minutes per pier at 2000×1500 pixels) and accurate capture of minute defects on rough surfaces (although the shooting time is extended to 25 minutes at 3500×2600 pixels, the recognition rate of minute cracks is still up to standard), thus meeting the efficiency and accuracy requirements of engineering inspection.

[0124] Understandably, the camera shooting parameters (resolution, texture filtering function) are adjusted according to the arithmetic mean deviation of roughness to adapt the shooting parameters to the roughness characteristics of the pier surface. When the surface is smooth, a conventional resolution is used to improve efficiency, while when the surface is rough, a high-definition resolution + texture filtering is used to ensure detail capture, balancing detection efficiency and defect recognition accuracy, and avoiding the loss of details or waste of resources caused by uniform parameters.

[0125] In some embodiments of this application, adjusting the camera position based on the flatness data includes:

[0126] When the flatness data shows that the height of the protrusion on the pier surface is greater than a preset protrusion height threshold and the depth of the depression on the pier surface is less than or equal to a preset depression depth threshold, the distance between the camera and the pier surface is increased by a first distance.

[0127] When the flatness data shows that the height of the protrusion on the pier surface is less than or equal to a preset protrusion height threshold and the depth of the depression on the pier surface is greater than a preset depression depth threshold, the distance between the camera and the pier surface is reduced by a second distance.

[0128] When the flatness data shows that the height of the protrusion on the pier surface is less than or equal to a preset protrusion height threshold and the depth of the depression on the pier surface is less than or equal to a preset depression depth threshold, the current position of the camera remains unchanged.

[0129] Specifically, surface protrusion height refers to the maximum vertical distance between the actual contour of the pier surface and the virtual reference surface (or the theoretical outer surface of the pier) within the detection area of ​​the pier surface, measured with reference to the continuous flat surface fitted within that area. Its core function is to reflect the degree of local outward protrusion of the pier surface, directly affecting whether the actual distance between the camera and the pier surface deviates from the optimal imaging distance, providing crucial data support for subsequent camera position adjustments. The preset protrusion height threshold is a critical value determined through statistical analysis of surface protrusion data of piers under various service conditions (new construction, 5 years of operation, 10 years of operation, etc.) and different environments (coastal, inland, high temperature, severe cold), combined with hardware parameters such as the focal length and resolution of the camera mounted on the pier-climbing robot. This threshold is determined by repeatedly testing the imaging clarity (e.g., whether a 0.2mm wide fine crack can be clearly identified) at different protrusion heights. When the actual surface protrusion height exceeds this threshold, the image captured by the camera will show blurred edges and loss of defect details, thus requiring camera position adjustments. Surface depression depth refers to the maximum vertical distance between the actual contour of the pier surface and the reference surface, measured with reference to the virtual reference surface. The absolute value of the vertical distance reflects the degree of inward concavity on the pier surface. Excessive concavity leads to a greater distance between the camera and the concave area, affecting image clarity and serving as a crucial basis for determining whether the camera needs to be moved closer. The preset concavity depth threshold and protrusion height threshold are determined using the same logic. They are critical values ​​determined by collecting a large amount of pier concavity sample data and combining it with camera imaging characteristic tests (such as the recognition accuracy of honeycomb textures at different concavity depths). Exceeding this threshold requires adjusting the camera to be closer to the pier surface. The first distance refers to the vertical distance between the camera and the pier surface when the surface protrusion height exceeds the preset protrusion height threshold. This distance must be balanced between image clarity and robot operation safety to avoid over-adjustment leading to collisions with other structures. The second distance refers to the vertical distance between the camera and the pier surface when the surface concavity depth exceeds the preset concavity depth threshold. This distance must be balanced to shorten the distance between the camera and the concave area to ensure image clarity. This distance must be balanced to ensure the camera does not collide with the edge of the concavity due to excessive proximity, while also meeting the clarity requirements for defect recognition.

[0130] Specifically, the surface protrusion height is obtained by using a ranging device (such as a laser rangefinder) mounted on a climbing robot to perform a high-density continuous scan of the pier surface detection area (e.g., collecting one data point per square centimeter). This acquires the three-dimensional coordinate data of all sampling points within the area. Then, data processing tools are used to remove obviously abnormal sampling points (such as extreme values ​​caused by surface deposits). A reference surface is fitted using the remaining continuous and flat sampling points. The vertical distance between each sampling point and the reference surface is calculated, and the maximum positive value (above the reference surface) is taken as the surface protrusion height of the detection area. The preset protrusion height threshold is obtained by first collecting data on different types (reinforced concrete, prestressed concrete) and different service conditions. A sample database was established using surface protrusion data of piers with varying ages. A simulation platform was then built in the laboratory to create pier surface models with protrusion heights gradually increasing from 0.1mm to 10mm. Each model had a pre-fabricated 0.2mm wide standard crack on its surface. Cameras were controlled to capture images at a fixed initial distance (e.g., 150mm, the optimal imaging distance for the camera). The clarity of crack identification in each image was analyzed, and the maximum protrusion height that could clearly identify the crack was recorded. Combined with the common range of actual pier protrusions in the sample database, the upper limit of the intersection of the two was taken as the preset protrusion height threshold. The acquisition of surface depression depth was similar to that of surface protrusion height; the three-dimensional coordinates of sampling points in the detection area were obtained by scanning with a ranging device, and a baseline was fitted. After aligning with the reference plane, the vertical distance between each sampling point and the reference plane is calculated. The maximum absolute value of the negative values ​​(below the reference plane) is taken as the surface indentation depth. The preset indentation depth threshold is obtained by collecting a large amount of pier indentation sample data, creating models with indentation depths ranging from 0.1mm to 10mm in the laboratory, and prefabricating standard honeycomb pitted defects in each model. The defect recognition accuracy is analyzed by taking pictures with a camera at the initial distance, recording the maximum indentation depth that can clearly identify the defect, and combining this with the actual pier indentation data distribution to determine the preset indentation depth threshold. The first distance is obtained in the laboratory by setting the surface protrusion height to the preset protrusion height threshold + 0.5mm (simulating a real scenario exceeding the threshold), and taking pictures... The initial camera distance is set as the optimal imaging distance. Then, the distance between the camera and the model surface is gradually increased (from 0.5mm to 5mm). After each adjustment, an image is captured and the clarity of the standard cracks in the raised area is analyzed. The value that can clearly identify the cracks and has the smallest adjustment distance is selected as the first distance. The second distance is obtained by setting the surface depression depth to a preset depression depth threshold + 0.5mm. The initial camera distance is set as the optimal imaging distance. The distance between the camera and the model surface is gradually decreased (from 0.5mm to 5mm). After each adjustment, an image is captured and the accuracy of the honeycomb texture in the depression area is analyzed. The value that can clearly identify the defects and has the smallest adjustment distance is selected as the second distance.

[0131] Specifically, in a defect detection project for an intercity railway bridge pier, when obtaining the surface protrusion height, a pier-climbing robot equipped with a laser rangefinder scanned the surface of a reinforced concrete pier that had been in operation for 6 years (selecting a detection area of ​​3-4m in height and 0.4m × 0.4m in area). The scanning density was 1 sampling point per square centimeter, obtaining the three-dimensional coordinates of 1600 sampling points. Two abnormal sampling points (coordinate deviation ±6mm) caused by surface laitance were removed using data processing software. A reference surface was fitted using 30 consecutive flat points from the remaining 1598 sampling points. The maximum vertical distance between each sampling point and the reference surface was calculated to be 3.5mm, which is the surface protrusion height of the detection area. A preset protrusion height threshold was then obtained. First, surface protrusion data of 40 railway piers with different operating years (2, 4, 6, and 8 years) along the railway line were collected. Statistics showed that the protrusion height was mostly distributed between 0.4 and 4.2 mm. Then, simulated pier surface models with protrusion heights of 0.5 mm, 1 mm, 2 mm, 3 mm, 4 mm, 4.5 mm, and 5 mm were created in the laboratory using concrete. Each model had a 0.2 mm wide crack pre-fabricated. The optimal imaging distance for the camera on the pier-climbing robot was 140 mm. After photographing each model, image analysis software was used to test the crack recognition rate. It was found that the crack recognition rate was over 95% when the protrusion height was no more than 4 mm, but dropped below 80% when it exceeded 4 mm. Based on the actual pier protrusion data distribution, a preset... The threshold for the protrusion height is 4mm. When obtaining the surface depression depth, the detection area of ​​the same pier with a height of 5-6m is scanned. After fitting the reference plane, the vertical distance between the minimum sampling point and the reference plane is calculated to be -3.2mm, and its absolute value of 3.2mm is the surface depression depth. When obtaining the preset depression depth threshold, the depression data of the above 40 piers are collected and statistically distributed between 0.3-3.8mm. Models with depression depths of 0.5mm, 1mm, 2mm, 3mm, 3.5mm, 4mm, and 4.5mm are made in the laboratory. Each model has a pre-made standard honeycomb hole with a diameter of 5mm. The camera is shot at an initial distance of 140mm. The honeycomb hole recognition rate is analyzed, and it is found that the recognition rate reaches 92% when the depression depth does not exceed 3.5mm. If the percentage is above 78%, it drops below 78%, and the preset indentation depth threshold is determined to be 3.5mm. When obtaining the first distance, a model with a protrusion height of 4.5mm (threshold 4mm + 0.5mm) was made in the laboratory. The initial camera distance was 140mm, and the distance was gradually increased to 141mm, 142mm, and 143mm. After shooting, the crack recognition rate was analyzed. It was found that when the distance increased by 2mm (i.e., 142mm), the crack recognition rate rebounded to 94%, and when it increased by 1mm, the recognition rate was only 88%. Therefore, the first distance was determined to be 2mm. When obtaining the second distance, a model with an indentation depth of 4mm (threshold 3.5mm + 0.5mm) was made. The initial camera distance was 140mm, and the distance was gradually decreased to 139mm and 138mm.Analysis of images taken at 5mm and 138mm distances revealed that the cell recognition rate rebounded to 93% when the distance decreased by 1.5mm (to 138.5mm), but dropped to only 87% when the distance decreased by 1mm. Therefore, the second distance was determined to be 1.5mm.

[0132] Understandably, adjusting the camera position based on the flatness data, and making targeted adjustments (moving away when there is a bulge and moving closer when there is a depression), ensures that the camera always maintains the optimal shooting distance from the pier surface, avoids blurry images due to uneven surfaces, further improves image quality, and provides distance adaptation guarantee for accurate defect identification.

[0133] In some embodiments of this application, when dividing the pier into detection areas for defect detection and adjusting the defect detection method based on the detection areas, the method includes:

[0134] Regarding the top detection area:

[0135] The edge enhancement mode of the camera on the climbing robot is activated, and the top detection area image is captured based on the shooting parameters and camera position adjusted in step S2 and using the edge enhancement mode.

[0136] Obtain the edge grayscale change rate of each edge in the top detection region image;

[0137] When the edge grayscale change rate is greater than the preset edge grayscale change rate threshold, it is determined that there are weathering cracks in the top detection area;

[0138] When the edge grayscale change rate is less than or equal to the preset edge grayscale change rate threshold, it is determined that there are no weathering cracks in the top detection area.

[0139] Specifically, the top inspection area refers to the transition area where the top of the bridge pier connects to the cap beam (or the beam body) and the specific range extending upwards from the top of the pier. This area, due to long-term exposure to the atmosphere, is susceptible to rain erosion, temperature changes, and ultraviolet radiation. Furthermore, it is affected by stress concentration caused by the load transfer from the cap beam, often exhibiting defects such as fine weathering cracks, surface concrete spalling, and leakage marks at the connection points. During inspection, special attention should be paid to the gaps at the interface with the cap beam and the damage to the edges of the pier top. The middle inspection area refers to the main central area of ​​the pier after removing the top and bottom. This area is the main load-bearing section of the pier and has a relatively intact structure, but it is susceptible to damage due to the concrete pouring process during construction (such as layered pouring). Factors such as excessively long construction intervals and insufficient compaction can easily lead to defects such as cold joints, honeycomb surfaces, and internal voids. Furthermore, due to the large height range in this area, the impact of the uniformity of concrete strength at different heights on defect detection must be considered. The bottom inspection area refers to the area of ​​the pier column near the ground (or the top surface of the foundation), near the waterline, or directly connected to the foundation structure. This area is susceptible to surface water immersion, soil erosion, and water flow scouring (if the pier column is located in water or in a rainy environment). It is also prone to the accumulation of mud, sand, debris, and other attachments, often resulting in deep concrete carbonation, surface corrosion, grooves formed by water scouring, and spalling defects. Surface attachments must be treated before inspection to avoid interfering with defect identification.

[0140] Specifically, when dividing the inspection area into top, middle, and bottom areas, the total height of the pier is first measured using a laser rangefinder mounted on the pier-climbing robot (the vertical distance from the top surface of the foundation cap or ground to the connection surface between the top of the pier and the cap beam), and the actual working parameters around the pier are recorded (such as ground elevation, normal water level, and bottom elevation of the cap beam). Then, using the lower edge of the connection surface between the top of the pier and the cap beam as a reference point, a preset vertical distance is measured downwards (this distance is determined based on the pier height and the area prone to top defects, usually 1 / 5-1 / 4 of the total pier height, or fixed at 0.5-2 meters, depending on the pier design dimensions). The area from this reference point to this measurement point is defined as the top inspection area; then, the area from the top surface of the foundation cap or ground to the bottom inspection area is measured using a laser rangefinder mounted on the pier-climbing robot. Using the ground as a reference point, measure a preset vertical distance upwards (this distance is determined based on the ground water level, the range of water erosion, or the depth of soil erosion; it is usually 1 / 5 to 1 / 4 of the total height of the pier, or fixed at 1-3 meters. If the pier is located in water, the area 0.5 meters above and 1 meter below the normal water level is used as the bottom core area). The area from this reference point to this measurement point is defined as the bottom detection area. Finally, the area between the lower edge of the top detection area on the pier and the upper edge of the bottom detection area is the middle detection area. If there are protruding structures on the pier surface (such as corbels) during the division process, the boundaries of adjacent detection areas need to be slightly adjusted according to the position and size of the protruding structures to ensure that each area does not overlap and completely covers the entire height range of the pier.

[0141] Specifically, taking a circular concrete pier of a highway bridge (designed total height 8 meters, of which the height from the top of the pier cap to the ground is 1 meter, the height of the pier above the ground is 7 meters, and the normal water level is 1 meter above the ground) as an example, when dividing the three inspection areas, firstly, the laser rangefinder mounted on the pier-climbing robot is used to measure upwards from the top of the pier cap to determine the actual height of the part of the pier above the ground as 7.2 meters (slightly higher than the design value due to construction error), and at the same time, the actual height of the normal water level is recorded as 1.1 meters above the ground; when delineating the top inspection area, the lower edge of the connection surface between the top of the pier and the cap beam is used as the reference point (the height of this reference point from the ground is determined to be 7.2 meters after distance measurement), and 1.5 meters is measured downwards (determined based on the stress concentration area range of this type of pier top and past defect statistics), then the top inspection area is 5 meters above the ground. The range is 7 meters to 7.2 meters. When delineating the bottom inspection area, considering that the pier is located in a rainy area, the ground is prone to water accumulation and the area near the normal water level is easily eroded, the ground (1 meter from the top surface of the pier cap) is used as the reference point to measure 1.6 meters upwards (covering the possible height of the ground water accumulation and the range of water erosion). Therefore, the bottom inspection area is the range from 0 meters to 1.6 meters above the ground (i.e., from the top surface of the pier cap to 1.6 meters above the ground). The middle inspection area is the range between the lower edge of the top inspection area (5.7 meters above the ground) and the upper edge of the bottom inspection area (1.6 meters above the ground), i.e., the area from 1.6 meters to 5.7 meters above the ground. During the division process, since the pier has no protruding structure, there is no need to adjust the area boundaries. In the end, the three areas completely cover the entire height of the pier above the ground, and the boundaries of each area are clear and do not overlap.

[0142] Specifically, the edge enhancement mode of the camera-equipped pier-climbing robot is an image optimization processing function specifically designed for bridge pier defect detection. Its core purpose is to highlight the contour features of fine structures by amplifying the grayscale differences of object edges in images captured by the camera. This mode first performs pixel-level analysis on the original image of the pier surface captured by the camera, calculating the grayscale difference between adjacent pixels. For edge areas with large differences (such as areas where cracks may exist), it amplifies the contrast by increasing the brightness of high grayscale pixels and decreasing the brightness of low grayscale pixels. At the same time, it suppresses redundant information in non-edge areas (such as smooth parts of the pier surface, slight textures, or stains) and reduces background interference. It is especially suitable for pier-climbing robots to inspect piers, addressing the problem of blurred contours of fine weathering cracks on the pier surface caused by uneven material and slight contamination. It can clearly separate the originally narrow and low grayscale contrast crack edges from the complex background, providing a basis for subsequent crack judgment. The presence of cracks provides clearer visual evidence; while weathering cracks are fine cracks caused by the long-term exposure of bridge piers to the natural environment, affected by factors such as wind, rain, temperature fluctuations, wet and dry cycles, and atmospheric pollutant erosion. These cracks result from physicochemical changes in the surface concrete of the pier (such as cement paste aging, separation of aggregate and cement paste interface, and uneven surface carbonization shrinkage). Typical characteristics include small width (mostly in the micrometer to millimeter range), shallow depth (usually only distributed within 5-50mm of the pier surface, rarely penetrating into the internal structure), relatively scattered distribution, and easy confusion with construction textures and minor scratches on the pier surface. Although they generally do not directly affect the overall structural safety of the pier in the early stages, their long-term existence can become a channel for moisture and corrosive substances to seep into the pier, accelerating the corrosion of internal steel bars and the deterioration of concrete. Therefore, pier climbing robots need to accurately capture these cracks through edge enhancement mode to avoid missed detection or misjudgment, and provide accurate basis for pier maintenance.

[0143] Specifically, the edge grayscale change rate refers to the rate at which the grayscale values ​​change between adjacent pixels in an image's edge region. It is typically calculated by dividing the difference in grayscale value between a pixel on the edge and its adjacent pixels by the pixel spacing (usually one pixel). It quantifies the abruptness of the grayscale transition from one region to another; a larger value indicates a more drastic grayscale change, and is more likely to be an edge formed by defects such as cracks. The preset edge grayscale change rate threshold is a critical value determined through extensive sample analysis. It is used to distinguish whether an edge in an image is caused by minor weathering cracks or normal surface texture. When the actual calculated edge grayscale change rate is greater than this threshold, it is determined that minor weathering cracks exist; otherwise, it is determined to be normal. On the surface, this threshold directly determines the sensitivity and accuracy of crack detection, and the need to reduce both missed and false detections must be considered. Obtaining the edge grayscale change rate requires first acquiring images of the top area of ​​the pier using the climbing robot's camera. After noise reduction (e.g., Gaussian filtering) and grayscale processing, edge detection algorithms (e.g., Canny operator, Sobel operator) are used to extract the edge contours in the image. Then, for each pixel on the contour, the difference in grayscale value between it and its adjacent pixels (e.g., horizontal, vertical, or diagonal directions) is calculated and divided by the physical distance between pixels (usually counted as 1 per pixel) to obtain the edge grayscale change rate of each point. Obtaining the preset edge grayscale change rate threshold requires collecting a large number of known images containing minute weathering cracks and those without cracks. For image samples of the top of a bridge pier with cracks, the grayscale change rate of all edges in both types of samples is calculated. Statistical analysis (such as plotting frequency distribution curves and calculating confusion matrices) is used to find a critical value that maximizes the distinction between crack edges and normal edges, ensuring the highest proportion of correctly identified cracks (true positive rate) while controlling the proportion of normal edges misidentified as cracks (false positive rate) within an acceptable range. This critical value is then determined as the preset threshold. In practical applications, such as inspecting the top of a bridge pier, 2000 sample images are first collected (1000 images manually confirmed to contain minor weathering cracks, and 1000 images confirmed to be crack-free). The images are preprocessed using OpenCV software (5×5 Gaussian kernel noise reduction, ...). (Grayscale conversion) The Sobel operator is used to extract edges (horizontal and vertical convolution kernels). For the extracted edge pixels, the grayscale difference between each pixel and its right-side neighbor is calculated (e.g., the grayscale of a pixel at a crack may suddenly change from 80 to 200, the difference is 120). Dividing this by the pixel spacing 1 gives the change rate of 120. Statistical analysis shows that the edge grayscale change rate of crack samples is mostly between 90-180, while that of crack-free samples is mostly between 20-70. Further ROC curve analysis shows that when the threshold is set to 85, 92% of cracks can be correctly identified and only 5% of normal edges are misidentified. Therefore, 85 is used as the preset edge grayscale change rate threshold for subsequent detection. When the calculated edge grayscale change rate is greater than 85, it is determined that there are minor weathering cracks.

[0144] Understandably, for the top detection area, the edge enhancement mode is combined with the edge grayscale change rate to determine weathering cracks. This focuses on weathering crack defects that are prone to appear at the top, strengthens crack features through edge enhancement, and then uses the grayscale change rate threshold for quantification to improve the targeting and accuracy of weathering crack identification in the top area and avoids confusion with defects in other areas.

[0145] In some embodiments of this application, when dividing the pier into detection areas for defect detection and adjusting the defect detection method based on the detection areas, the method further includes:

[0146] Regarding the central detection area:

[0147] Activate the detail enhancement mode of the camera mounted on the climbing robot, and capture an image of the central detection area based on the shooting parameters and camera position adjusted in step S2 and using the detail enhancement mode.

[0148] The image of the central detection region is divided into several grid units using a grid partitioning comparison method;

[0149] When the gray value of a pixel in a certain grid cell is less than the first threshold and the proportion of the number of pixels with a gray value less than the first threshold to the total number of pixels in the current grid cell is greater than the first proportion, it is determined that the current grid cell has a honeycomb surface and the current grid is recorded as the first grid cell.

[0150] When the number of pixels with gray values ​​greater than or equal to the first threshold or the number of pixels with gray values ​​less than the first threshold in a certain grid cell is less than or equal to the first proportion of the total number of pixels in the current grid cell, it is determined that the current grid cell does not have a honeycomb surface and the current grid is recorded as the second grid cell.

[0151] When the proportion of the number of the first grid unit to the total number of grid units in the central detection area is greater than a preset regional defect proportion threshold and the proportion of the number of the second grid unit to the total number of grid units in the central detection area is less than or equal to the preset regional defect proportion threshold, it is determined that there is honeycomb pitting in the central detection area.

[0152] When the proportion of the number of the second grid cells to the total number of grid cells in the central detection area is greater than a preset regional defect proportion threshold and the proportion of the number of the first grid cells to the total number of grid cells in the central detection area is less than or equal to a preset regional defect proportion threshold, it is determined that there is no honeycomb surface in the central detection area.

[0153] Specifically, the "detail enhancement mode" is a specific imaging mode for the pier-climbing robot equipped with a camera, targeting the central inspection area of ​​bridge piers. It uses algorithms to enhance the grayscale differences in the subtle textures of the concrete surface, highlighting details such as material density variations and minute bumps, providing a clearer image foundation for subsequent defect identification. This is particularly suitable for highlighting surface defects easily confused with the background, such as construction cold joints and honeycomb pitting. The "grid partitioning comparison method" is a systematic approach to processing the image of the central inspection area. It divides the image into several independent grid units according to a preset size (determined by combining typical defect sizes with camera resolution). Defects are determined by analyzing the pixel grayscale distribution characteristics within each unit. This method reduces the impact of overall image complexity on detection accuracy and enables defect localization. The classification includes: "Construction cold joints" are hidden or visible gaps formed during concrete pouring due to excessively long intervals between pouring batches, resulting in poor bonding between newly poured concrete and partially set old concrete. In images, they often appear as abnormal grayscale differences in local pixels within a grid cell (within the first range). Essentially, they represent weak areas of concrete bonding. "Honeycomb pitting" is a defect on the concrete surface caused by inadequate compaction, failure to expel air bubbles in time, or loss of cement slurry. "Honeycomb" is characterized by numerous irregular holes in a localized area (corresponding to a proportion of low grayscale pixels exceeding the first proportion within the grid cell), while "pitting" is characterized by a rough surface with dense small pits (corresponding to some pixels within the grid cell having grayscale values ​​below the first threshold). Both can affect the durability and appearance quality of concrete structures.

[0154] Specifically, "pixel grayscale difference value" refers to the difference between the maximum and minimum values ​​of all pixel grayscale values ​​within the same grid cell of the image in the central detection area, reflecting the degree of color variation of the concrete surface in that area; "first range" is a pre-defined range of pixel grayscale difference values ​​used to determine the presence of construction cold joints. When the pixel grayscale difference value of a certain grid cell falls into this range, it indicates that the area may have cold joints due to pouring intervals during construction; "pixel grayscale value" is the brightness value of a single pixel in the image, typically ranging from 0 to 255 (0 being pure black and 255 being pure white), used to quantify the brightness of the concrete surface; "first threshold" is the grayscale threshold value used to distinguish between normal concrete areas and honeycomb-like pitted areas. The pixels with this value typically correspond to areas with surface defects such as holes and looseness. The "first proportion" is the minimum percentage of pixels with gray values ​​below the first threshold within a grid cell when honeycomb surface defects are identified. When this percentage is reached or exceeded, it indicates the presence of honeycomb surface defects in that area. The "second range" is a pre-defined range of pixel gray value differences used to determine whether there are no construction cold joints or honeycomb surface defects. When the pixel gray value difference of a grid cell falls within this range and all pixel gray values ​​are greater than the first threshold, it indicates that the concrete surface in that area is smooth and without obvious defects. These parameters are obtained through the following specific operations: First, a large number of samples from the central area of ​​bridge piers containing construction cold joints, honeycomb surface defects, and normal surfaces are collected. The image is processed, and defective and normal areas are manually marked. Then, for the marked cold joint areas, the pixel grayscale difference value within each grid cell is calculated, and the distribution range of these differences is statistically analyzed. The interval that can stably distinguish between cold joints and normal areas is taken as the first range. For normal surface areas, the distribution of their pixel grayscale values ​​is statistically analyzed, and the critical value that can distinguish between normal and defective areas is taken as the first threshold. For the marked honeycomb-patterned areas, the proportion of pixels with grayscale values ​​below the first threshold in each grid cell is calculated, and the distribution of these proportions is statistically analyzed. The minimum proportion that can stably determine honeycomb-patterned areas is taken as the first percentage. For normal surface areas, the distribution range of their pixel grayscale difference values ​​is calculated, and the interval that can stably distinguish between cold joints and normal areas is taken as the first percentage. The area without cold joints is designated as the second range. Finally, these parameters are verified and adjusted through multiple experiments to ensure accurate defect identification. In practical applications, for example, in the inspection of the middle part of a bridge pier, 1000 images containing construction cold joints, honeycomb pitting, and normal surfaces are collected first, marking 500 construction cold joint areas, 300 honeycomb pitting areas, and 200 normal areas. For cold joint areas, the image is divided into 100×100 pixel grid units, and the pixel grayscale difference value of each cold joint grid is calculated. It is found that it is mostly between 30 and 80, so the first range is set to 30-80. For normal areas, the pixel grayscale values ​​are mostly between 100 and 200, and 90 is taken as the first threshold (pixels below 90 mostly correspond to defects).For the honeycomb-patterned area, the percentage of pixels with a grayscale value below 90 in each grid was calculated. It was found that when the percentage exceeded 30%, it could be stably identified as a honeycomb-patterned area. Therefore, the first percentage was set to 30%. For the grid cells in the normal area, the pixel grayscale difference value was calculated. It was found that it was mostly between 0 and 20. Therefore, the second range was set to 0-20. This completed the determination of these parameters.

[0155] Specifically, the "regional defect proportion threshold" refers to the core quantitative indicator used to determine whether honeycomb surface defects exist in the central inspection area of ​​a pier. Specifically, it is the critical proportion of the number of first grid cells with honeycomb surface defects in the central inspection area to the total number of grid cells in the area. The setting of this threshold needs to be combined with the structural functional requirements of the central pier, the degree of impact of honeycomb surface defects on concrete strength and durability, and the requirements of the bridge technical condition level for the proportion of minor component defects in JTG5120-2021 "Technical Specification for Maintenance of Highway Bridges and Culverts" (e.g., Class 2 bridges have less than 10% moderate defects in minor components, and Class 3 bridges have 10% to 20% severe defects in minor components) and Table 5.1 of JTGTH21-2011 "Standard for Evaluation of Technical Condition of Highway Bridges".The evaluation criteria for honeycomb pitting in 1-1 (scale 2: cumulative area ≤ 50% of component area; scale 3: cumulative area > 50% of component area) aim to avoid misjudging the entire area as defective due to the presence of defects in a small number of local grid units, while also preventing the omission of safety hazards due to an excessively high proportion of defective grid units. It serves as a crucial bridge connecting the assessment of individual grid unit defects with the assessment of overall regional defects, ensuring that the assessment of honeycomb pitting defects in the central inspection area is both scientific and engineering-practical. This assessment requires multi-step, multi-dimensional testing and statistical analysis: firstly, data on different service years, different concrete strength grades, and other factors are collected. For pier samples with the same construction process and known presence of honeycomb and pitted defects, common working conditions were ensured. Secondly, each sample was divided into several grid units according to the actual grid zoning method used in the testing. Each grid unit was manually checked and marked with high-precision testing equipment to determine if honeycomb and pitted defects were actually present. The percentage of defective grid units in each sample was then statistically analyzed. Next, the statistical data was analyzed according to the relevant requirements of the two standards mentioned above. Percentile and cluster analysis methods were used to determine the critical proportion that could cover the vast majority of real-area honeycomb and pitted samples while excluding a small number of locally defective samples. Finally, this value was verified and optimized on samples under different working conditions to ensure accurate differentiation of areas. The presence of honeycomb and pitted surfaces was investigated. For example, in a highway bridge pier inspection project, 100 piers in different service conditions were selected (including 20 newly built piers without honeycomb and pitted surfaces, 30 piers with minor honeycomb and pitted surfaces after 3-5 years of operation, 30 piers with obvious honeycomb and pitted surfaces after 5-10 years of operation, and 20 piers with honeycomb and pitted surfaces before repair). The central inspection area was divided into 20mm×20mm grid units. Through manual observation combined with high-definition camera marking of defect grids, it was found that the proportion of defect grids in newly built piers was less than 5%, the proportion of piers with minor honeycomb and pitted surfaces was 8%-12% (corresponding to category 2 of JTGTH21-2011 scale), and the proportion of piers with obvious honeycomb and pitted surfaces was... The proportion of defective components was 15%-30% (corresponding to the proportion of minor component defects in bridges of scale 3 in JTGTH21-2011 and JTG5120-2021), and the proportion of defective piers before repair was 18%-25%. Referring to the requirements for repair and maintenance of bridges of type 3 in JTG5120-2021, it was found that when the proportion of defective mesh exceeded 10%, more than 78% of the samples showed potential durability risks after structural calculation. Using 10% as the initial threshold, verification was conducted on 50 piers with unknown defects. This accurately identified 90% of the actual regional honeycomb-like surface defects in the piers without misjudging local defects. Therefore, 10% was ultimately determined as the "regional defect proportion threshold" for this project and similar working conditions.

[0156] Understandably, for the central inspection area, a detail enhancement mode combined with grid partitioning is used to inspect the honeycomb surface. Detail enhancement highlights the grayscale differences of the honeycomb surface, while grid partitioning makes the judgment more precise. Furthermore, the presence of defects in the area is determined by the grid ratio, thereby improving the accuracy and reliability of the central honeycomb surface inspection.

[0157] In some embodiments of this application, when dividing the pier into detection areas for defect detection and adjusting the defect detection method based on the detection areas, the method further includes:

[0158] Regarding the bottom detection area:

[0159] The contour enhancement mode of the camera on the climbing robot is activated, and the bottom detection area image is captured based on the shooting parameters and camera position adjusted in step S2 and the contour enhancement mode is used.

[0160] The percentage of the area covered by the surface deposits in the bottom detection area relative to the total area of ​​the bottom detection area;

[0161] When the coverage area ratio is greater than the second ratio, the surface attachments of the bottom detection area are cleaned and the image of the bottom detection area is re-captured;

[0162] When the coverage area ratio is less than or equal to the second ratio, the defect features present in the bottom detection area are determined based on the area of ​​the continuous blank area.

[0163] When the area of ​​a continuous blank region in the bottom detection region image is greater than the first area threshold, it is determined that there is a peeling defect in the bottom detection region.

[0164] When the area of ​​a continuous blank region in the bottom detection region image is less than the first area threshold, it is determined that there is no peeling defect in the bottom detection region.

[0165] Specifically, the contour enhancement mode is an image capture assistance function specifically activated by the camera on the pier-climbing robot for the bottom detection area. It enhances the contour boundary features between different areas by highlighting areas where pixel grayscale values ​​change abruptly in the image (such as the boundary between intact concrete and attached materials like mud, moss, etc., on the pier surface, or the transition area between intact concrete and the edge of spalling defects). This avoids confusion between impurities attached to the bottom detection area due to long-term water flow and the pier body or spalling defects, ensuring clear distinction of the physical boundaries of each area during subsequent image analysis, laying the foundation for accurate identification of spalling defects. The continuous blank area refers to a continuous area within a set pixel range in the bottom detection area image processed by the contour enhancement mode, exhibiting a significant difference in pixel grayscale values ​​from the surrounding intact concrete area. This area is formed because the exposed base material (such as reinforcing steel or concrete pad) after the concrete surface at the bottom of the pier has detached, and its optical reflection characteristics differ from the intact concrete surface. In the image, this appears as a continuous area with a clear boundary between its grayscale values ​​and the surrounding normal pixels. Furthermore, it must meet the characteristics of "continuous" (no scattered normal concrete pixels interrupting the area) and "patchy" (the area reaches the set minimum recognition threshold) to distinguish it from isolated gray-scale anomalies formed by scattered attachments or tiny impurities. Peeling defects are common structural defects at the bottom of bridge piers. They are mainly caused by factors such as long-term water erosion that carries away concrete particles from the pier surface, chemical erosion in the environment leading to a decrease in the bonding strength of the concrete surface layer, freeze-thaw cycles damaging the concrete surface structure, or insufficient strength of the pier concrete itself. It manifests as the concrete surface layer of the pier peeling off in sheets or blocks from the body, forming a recessed or exposed base layer area. In the detection of this defect, it needs to be identified by continuous blank areas in the image captured by contour enhancement mode. Its core difference from honeycomb pitting (only uneven surface, no concrete surface peeling) and weathering cracks (only linear cracks, no surface peeling) is that there is substantial peeling of the concrete surface layer, and the peeling area is presented in the form of continuous blank areas in the image. It is a type of defect that needs to be identified in the bottom detection area.

[0166] Specifically, the second percentage is a critical ratio used to determine whether surface attachments in the bottom inspection area of ​​the pier need to be cleaned. Specifically, it refers to the maximum proportion of the area covered by surface attachments in the bottom inspection area to the total area of ​​the bottom inspection area. When the proportion of attachment coverage does not exceed this ratio, no cleaning is needed, and subsequent image analysis can accurately identify the peeling defects in the bottom inspection area. If it exceeds this ratio, the attachments will obscure or interfere with the characteristics of the peeling defects, leading to inaccurate defect identification; therefore, the attachments must be cleaned first. The first area threshold is a critical area value used to determine whether there are peeling defects caused by water erosion in the bottom inspection area of ​​the pier. Specifically, it refers to the minimum area of ​​a continuous blank area in the bottom inspection area image. When the area of ​​a continuous blank area in the image exceeds this threshold, it can be determined that the blank area is a peeling defect on the pier surface caused by water erosion. If it does not exceed this threshold, the blank area may be surface stains, minor wear, etc. The normal surface features of non-stripping defects are used to avoid misjudging non-defect features as stripping defects. To obtain the second proportion, samples of the bottom of piers under different bridges, different service years, and different environments (such as near water, with more soil coverage, etc.) are collected first. The total area of ​​the bottom detection area and the coverage area of ​​surface attachments (such as mud, moss, scale, etc.) of each sample are measured, and the proportion of attachment coverage area of ​​each sample is calculated. Then, for each sample with a coverage area proportion, images are captured using contour enhancement mode based on the shooting parameters and camera position determined in step S2, and stripping defects are identified. The identification accuracy of stripping defects under different coverage area proportions is recorded. Finally, the identification accuracy data is statistically analyzed, and the proportion of attachment coverage area corresponding to when the identification accuracy begins to decrease significantly (such as the accuracy rate dropping from above 95% to below 85%) is selected as the second proportion to ensure that high defect identification accuracy can be achieved without cleaning under this proportion.To obtain the first area threshold, artificial simulated pier spalling defect specimens of different sizes (e.g., 2cm², 5cm², 8cm², 10cm², 15cm², etc.) need to be made. Multiple specimens of each size are made and installed at the bottom of simulated piers with the same material and surface condition as the actual piers. Then, based on the shooting parameters and camera position determined in step S2, images of each simulated specimen are captured using contour enhancement mode, and the area of ​​continuous blank regions in the images corresponding to different simulated defect areas is recorded. Then, spalling defect judgment tests are performed on each continuous blank region area, and the defect misjudgment rate (the proportion of non-defects misjudged as defects or missed defects) under different blank region areas is calculated. Finally, the area of ​​continuous blank regions with the lowest misjudgment rate (e.g., misjudgment rate below 5%) that can effectively distinguish between real spalling defects and minor surface imperfections is selected as the first area threshold. In practical applications, when obtaining the second proportion, 10 water-facing piers of a cross-river bridge can be selected as samples. First, the detection area at the bottom of each pier (set as diameter) is measured. The total area of ​​a 1m circular region (approximately 0.785m²) was determined. The area covered by the attached material was calculated by manually cleaning and weighing the material, combined with the density of the material, to determine the coverage area of ​​the attached material at the bottom of each pier. The calculated coverage areas of the 10 piers were 15%, 22%, 28%, 30%, 35%, 40%, 45%, 50%, 55%, and 60%, respectively. Then, based on the concrete strength grade and aggregate type of the bridge piers, the shooting parameters (such as resolution 1920×1080, brightness 500cd / m², contrast 1:2) and the camera position (30cm from the pier surface) were determined. Contour enhancement mode was used to capture images of the bottom of each pier and identify spalling defects. It was found that when the coverage area was ≤30%, the actual spalling defects were accurately identified 9 out of 10 times (accuracy rate above 90%). When the coverage area was >30%, the accuracy rate dropped below 70% (e.g., only 6 accurate identifications at 35%). Therefore, the second coverage area in this scenario was determined to be 30%.When obtaining the first area threshold, a simulated pier can be created using C30 concrete with the same surface roughness as the bridge pier. Three spalling defects with areas of 5cm², 8cm², 10cm², 12cm², and 15cm² are manually chiseled into the bottom of the simulated pier, for a total of 15 simulated defects. Images of each simulated defect are captured using the same shooting parameters and camera position. The areas of consecutive blank areas in the images correspond to approximately 4.8cm², 7.9cm², 9.8cm², 11.7cm², and 15cm², respectively. 4.5cm²; Judgment tests were conducted on these blank areas, and it was found that when the area of ​​the blank area was ≥10cm² (corresponding to the actual simulated defect area of ​​10cm²), there was only 1 false judgment in 15 tests (false judgment rate of 6.7%), and it was able to eliminate the false judgment of surface scratches with an area <8cm² (corresponding to the blank area of ​​the image <8cm²). When the area of ​​the blank area was <10cm², the false judgment rate rose to more than 15% (e.g., 3 false judgments in an 8cm² blank area). Therefore, the first area threshold in this scenario was determined to be 10cm².

[0167] Understandably, for the bottom detection area, the contour enhancement mode is combined with the removal of attachments and the determination of peeling defects in blank areas. First, the attachments are removed to eliminate interference, and then contour enhancement and area threshold determination are used to accurately identify peeling defects at the bottom that are easily caused by water erosion, thereby improving the effectiveness of bottom area defect detection.

[0168] In some embodiments of this application, when determining the defect level of a bridge pier based on the defect characteristics of each of the detection areas and generating bridge pier defect detection results, the process includes:

[0169] When the weathering cracks are present in the top detection area, the honeycomb surface is present in the middle detection area, and the spalling defects are present in the bottom detection area, the bridge pier defect level is determined to be Level 1.

[0170] When the weathering cracks are present in the top detection area, the honeycomb surface is present in the middle detection area, and the spalling defects are not present in the bottom detection area, the defect level of the bridge pier column is determined to be Level II.

[0171] When the weathering cracks are present in the top detection area, the honeycomb surface is not present in the middle detection area, and the peeling defects are present in the bottom detection area, the defect level of the bridge pier is determined to be level three.

[0172] When the top detection area does not have weathering cracks, the middle detection area has honeycomb pitting, and the bottom detection area has spalling defects, the bridge pier defect level is determined to be level four.

[0173] When the top detection area does not have weathering cracks, the middle detection area does not have honeycomb pitting, and the bottom detection area has peeling defects, the bridge pier defect level is determined to be level five.

[0174] When the top detection area does not have weathering cracks, the middle detection area has honeycomb pitting, and the bottom detection area does not have spalling defects, the bridge pier defect level is determined to be level six.

[0175] When the weathering cracks are present in the top detection area, the honeycomb surface is not present in the middle detection area, and the peeling defects are not present in the bottom detection area, the defect level of the bridge pier is determined to be level seven.

[0176] When the top detection area does not have weathering cracks, the middle detection area does not have honeycomb pitting, and the bottom detection area does not have spalling defects, the bridge pier defect level is determined to be level eight.

[0177] The defect levels and the defect features of each of the detection areas are integrated to generate a bridge pier defect detection result that includes the defect levels and the defect features of each of the detection areas.

[0178] Specifically, Level 1 defects correspond to weathering cracks in the top inspection area, honeycomb surface defects in the middle inspection area, and spalling defects in the bottom inspection area. This combination of defects covers the entire inspection area of ​​the pier, having the most significant impact on the pier's load-bearing capacity, durability, and structural stability, resulting in the highest structural safety risk. Level 2 defects correspond to weathering cracks in the top inspection area, honeycomb surface defects in the middle inspection area, but no spalling defects in the bottom inspection area. Compared to Level 1, this reduces the impact of bottom spalling defects, eliminating the influence on the structural integrity of the pier's bottom, and thus lowering the overall structural risk. Level 3 defects correspond to weathering cracks in the top inspection area, no honeycomb surface defects in the middle inspection area, but spalling defects in the bottom inspection area. Compared to Level 1, this reduces the impact of honeycomb surface defects in the middle, eliminating the influence on the concrete density in the middle of the pier, and thus lowering the overall structural risk. Level 4 defects correspond to no weathering cracks in the top inspection area, honeycomb surface defects in the middle inspection area, and spalling defects in the bottom inspection area. Compared to Level 1, this reduces the impact of top weathering cracks, eliminating the impact on the pier's top weathering resistance, and thus lowering the overall structural risk. The structural risk is lower than Level 1; Level 5 defect level corresponds to the absence of weathering cracks in the top inspection area, the absence of honeycomb pitting in the middle inspection area, but the presence of spalling defects in the bottom inspection area. Spalling defects are only present at the bottom, mainly affecting the structural stability of the connection between the pier bottom and the foundation. The overall structural risk is lower than Level 4. Level 6 defect level corresponds to the absence of weathering cracks in the top inspection area, the presence of honeycomb pitting in the middle inspection area, but the absence of spalling defects in the bottom inspection area. Honeycomb pitting defects are only present in the middle, mainly affecting the density and impermeability of the concrete in the middle of the pier. The overall structural risk is lower than Level 4. Level 7 defect level corresponds to the presence of weathering cracks in the top inspection area, the absence of honeycomb pitting in the middle inspection area, and the absence of spalling defects in the bottom inspection area. Weathering cracks are only present at the top, mainly affecting the weathering resistance of the top of the pier. The overall structural risk is lower than Level 6. Level 8 defect level corresponds to the absence of weathering cracks in the top inspection area, the absence of honeycomb pitting in the middle inspection area, and the absence of spalling defects in the bottom inspection area. All inspection areas of the pier are free of defects, indicating good structural integrity, density, and stability, with the lowest structural safety risk.

[0179] Understandably, by identifying defect levels based on the characteristics of defects in each region and integrating the generated results, a unified and quantitative standard for determining defect levels can be established to avoid the subjectivity of human experience. At the same time, standardized results containing both levels and defect characteristics can be generated, providing a clear and reliable basis for the safety assessment of bridge piers and piers, and facilitating subsequent maintenance decisions.

[0180] Reference Figure 2 As shown in some embodiments of this application, a rapid defect detection system for bridge piers using a pier-climbing robot includes:

[0181] The camera brightness adjustment module is used to obtain the surface hardness and carbonation depth of the pier column, obtain the concrete strength grade and aggregate type of the pier column based on the surface hardness and carbonation depth, and adjust the brightness and contrast of the camera mounted on the pier climbing robot based on the concrete strength grade and aggregate type.

[0182] The camera position adjustment module is used to acquire the arithmetic mean deviation of the roughness and the flatness data of the pier surface, and adjust the shooting parameters of the camera based on the arithmetic mean deviation of the roughness, and adjust the position of the camera based on the flatness data.

[0183] The regional defect detection module is used to divide the pier into detection areas for defect detection, and adjust the defect detection method based on the detection areas. The detection areas include: a top detection area, a middle detection area, and a bottom detection area.

[0184] The defect result generation module is used to determine the defect level of the bridge pier based on the defect characteristics of each detection area, and generate the bridge pier defect detection results.

[0185] It is understandable that the above-mentioned method and system for rapid detection of bridge pier defects using a climbing robot have the same beneficial effects, and will not be elaborated further here.

[0186] It should be noted that:

[0187] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. In some instances, well-known structures and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0188] Furthermore, those skilled in the art will understand that although some embodiments described herein include certain features included in other embodiments but not others, combinations of features from different embodiments are meant to be within the scope of this application and form different embodiments.

[0189] The above description is merely a preferred embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A rapid detection method for bridge pier defects using a pier-climbing robot, characterized in that, include: S1, obtain the surface hardness and carbonation depth of the pier column, obtain the concrete strength grade and aggregate type of the pier column based on the surface hardness and carbonation depth, and adjust the brightness and contrast of the camera mounted on the pier climbing robot based on the concrete strength grade and aggregate type. S2, acquire the arithmetic mean deviation of roughness and flatness data of the pier surface, and adjust the shooting parameters of the camera based on the arithmetic mean deviation of roughness, and adjust the position of the camera based on the flatness data; S3, divide the pier into inspection areas for defect detection, and adjust the defect detection method based on the inspection areas. The inspection areas include: top inspection area, middle inspection area and bottom inspection area; S4, determine the defect level of the bridge pier based on the defect characteristics of each detection area, and generate the bridge pier defect detection results; When the top detection area has weathering cracks, the middle detection area has honeycomb pitting, and the bottom detection area has spalling defects, the bridge pier defect level is determined to be Level 1. When the top detection area has weathering cracks, the middle detection area has honeycomb pitting, and the bottom detection area has no spalling defects, the bridge pier defect level is determined to be Level II. When the top detection area has weathering cracks, the middle detection area does not have honeycomb pitting, and the bottom detection area has peeling defects, the bridge pier defect level is determined to be level three. When there are no weathering cracks in the top detection area, honeycomb pitting in the middle detection area, and peeling defects in the bottom detection area, the bridge pier defect level is determined to be level four. When there are no weathering cracks in the top detection area, no honeycomb pitting in the middle detection area, and peeling defects in the bottom detection area, the bridge pier defect level is determined to be level five. When there are no weathering cracks in the top detection area, honeycomb pitting in the middle detection area, and no spalling defects in the bottom detection area, the bridge pier defect level is determined to be level six. When the top detection area has weathering cracks, the middle detection area does not have honeycomb pitting, and the bottom detection area does not have spalling defects, the bridge pier defect level is determined to be level seven. When there are no weathering cracks in the top detection area, no honeycomb pitting in the middle detection area, and no spalling defects in the bottom detection area, the bridge pier defect level is determined to be level eight. The defect levels and the defect features of each of the detection areas are integrated to generate a bridge pier defect detection result that includes the defect levels and the defect features of each of the detection areas.

2. The rapid detection method for bridge pier defects using a pier-climbing robot according to claim 1, characterized in that, When determining the concrete strength grade and aggregate type of the pier column based on the surface hardness and carbonation depth, the following are included: When the surface hardness is within a first range and the carbonation depth is within a first interval, the concrete strength grade is determined to be the first grade and the aggregate type is the first type. When the surface hardness is within a first range and the carbonation depth is within a second range, the concrete strength grade is determined to be the second grade and the aggregate type is the second type. When the surface hardness is in the first range and the carbonation depth is in the third range, the concrete strength grade is determined to be the third grade and the aggregate type is the third type. When the surface hardness is in the second range and the carbonation depth is in the first range, the concrete strength grade is determined to be the fourth grade and the aggregate type is the fourth type. When the surface hardness is within the second range and the carbonation depth is within the second interval, the concrete strength grade is determined to be the fifth grade and the aggregate type is the fifth type. When the surface hardness is in the second range and the carbonation depth is in the third range, the concrete strength grade is determined to be grade six and the aggregate type is type six. When the surface hardness is in the third range and the carbonation depth is in the first range, the concrete strength grade is determined to be grade seven and the aggregate type is type seven. When the surface hardness is in the third range and the carbonation depth is in the second range, the concrete strength grade is determined to be the eighth grade and the aggregate type is the eighth type. When the surface hardness is in the third range and the carbonation depth is in the third interval, the concrete strength grade is determined to be the ninth grade and the aggregate type is the ninth type.

3. The rapid detection method for bridge pier defects using a pier-climbing robot according to claim 2, characterized in that, When adjusting the brightness and contrast of the camera mounted on the climbing robot based on the concrete strength grade and the aggregate type, the following steps are included: When the concrete strength grade is first grade and the aggregate type is first type, the brightness is adjusted to a first value and the contrast is adjusted to a first ratio. When the concrete strength grade is the second grade and the aggregate type is the second type, the brightness is adjusted to the second value and the contrast is adjusted to the second ratio. When the concrete strength grade is grade three and the aggregate type is type three, the brightness is adjusted to the third value and the contrast is adjusted to the third ratio. When the concrete strength grade is fourth grade and the aggregate type is fourth type, the brightness is adjusted to the fourth value and the contrast is adjusted to the fourth ratio. When the concrete strength grade is grade 5 and the aggregate type is type 5, the brightness is adjusted to the fifth value and the contrast is adjusted to the fifth ratio. When the concrete strength grade is grade 6 and the aggregate type is type 6, the brightness is adjusted to the sixth value and the contrast is adjusted to the sixth ratio. When the concrete strength grade is grade 7 and the aggregate type is type 7, the brightness is adjusted to the seventh value and the contrast is adjusted to the seventh ratio. When the concrete strength grade is grade 8 and the aggregate type is type 8, the brightness is adjusted to the eighth value and the contrast is adjusted to the eighth ratio. When the concrete strength grade is grade nine and the aggregate type is type nine, the brightness is adjusted to the ninth value and the contrast is adjusted to the ninth ratio.

4. The rapid detection method for bridge pier defects using a pier-climbing robot according to claim 3, characterized in that, When adjusting the camera's shooting parameters based on the roughness arithmetic mean deviation, the following steps are included: When the roughness arithmetic mean deviation is less than or equal to the preset roughness judgment threshold, the camera is controlled to take pictures at a resolution of the first horizontal pixel number multiplied by the first vertical pixel number, and the texture filtering function is turned off. When the roughness arithmetic mean deviation is greater than the preset roughness judgment threshold, the camera is controlled to take pictures at a resolution of the second horizontal pixel number multiplied by the second vertical pixel number, and the texture filtering function is enabled. The number of second horizontal pixels is greater than the number of first horizontal pixels, and the number of second vertical pixels is greater than the number of first vertical pixels.

5. A rapid detection method for bridge pier defects using a pier-climbing robot according to claim 4, characterized in that, Adjusting the camera position based on the flatness data includes: When the flatness data shows that the height of the protrusion on the pier surface is greater than a preset protrusion height threshold and the depth of the depression on the pier surface is less than or equal to a preset depression depth threshold, the distance between the camera and the pier surface is increased by a first distance. When the flatness data shows that the height of the protrusion on the pier surface is less than or equal to a preset protrusion height threshold and the depth of the depression on the pier surface is greater than a preset depression depth threshold, the distance between the camera and the pier surface is reduced by a second distance. When the flatness data shows that the height of the protrusion on the pier surface is less than or equal to a preset protrusion height threshold and the depth of the depression on the pier surface is less than or equal to a preset depression depth threshold, the current position of the camera remains unchanged.

6. A rapid detection method for bridge pier defects using a pier-climbing robot according to claim 5, characterized in that, When dividing the pier into inspection areas for defect detection, and adjusting the defect detection method based on the inspection areas, the process includes: Regarding the top detection area: The edge enhancement mode of the camera on the climbing robot is activated, and the top detection area image is captured based on the shooting parameters and camera position adjusted in step S2 and using the edge enhancement mode. Obtain the edge grayscale change rate of each edge in the top detection region image; When the edge grayscale change rate is greater than the preset edge grayscale change rate threshold, it is determined that there are weathering cracks in the top detection area; When the edge grayscale change rate is less than or equal to the preset edge grayscale change rate threshold, it is determined that there are no weathering cracks in the top detection area.

7. A rapid detection method for bridge pier defects using a pier-climbing robot according to claim 6, characterized in that, When dividing the pier into inspection areas for defect detection, and adjusting the defect detection method based on the inspection areas, the method further includes: Regarding the central detection area: Activate the detail enhancement mode of the camera mounted on the climbing robot, and capture an image of the central detection area based on the shooting parameters and camera position adjusted in step S2 and using the detail enhancement mode. The image of the central detection region is divided into several grid units using a grid partitioning comparison method; When the gray value of a pixel in a certain grid cell is less than the first threshold and the proportion of the number of pixels with a gray value less than the first threshold to the total number of pixels in the current grid cell is greater than the first proportion, it is determined that the current grid cell has a honeycomb surface and the current grid is recorded as the first grid cell. When the number of pixels with gray values ​​greater than or equal to the first threshold or the number of pixels with gray values ​​less than the first threshold in a certain grid cell is less than or equal to the first proportion of the total number of pixels in the current grid cell, it is determined that the current grid cell does not have a honeycomb surface and the current grid is recorded as the second grid cell. When the proportion of the number of the first grid unit to the total number of grid units in the central detection area is greater than a preset regional defect proportion threshold and the proportion of the number of the second grid unit to the total number of grid units in the central detection area is less than or equal to the preset regional defect proportion threshold, it is determined that there is honeycomb pitting in the central detection area. When the proportion of the number of the second grid cells to the total number of grid cells in the central detection area is greater than a preset regional defect proportion threshold and the proportion of the number of the first grid cells to the total number of grid cells in the central detection area is less than or equal to a preset regional defect proportion threshold, it is determined that there is no honeycomb surface in the central detection area.

8. A rapid detection method for bridge pier defects using a pier-climbing robot according to claim 7, characterized in that, When dividing the pier into inspection areas for defect detection, and adjusting the defect detection method based on the inspection areas, the method further includes: Regarding the bottom detection area: The contour enhancement mode of the camera on the climbing robot is activated, and the bottom detection area image is captured based on the shooting parameters and camera position adjusted in step S2 and the contour enhancement mode is used. The percentage of the area covered by the surface deposits in the bottom detection area relative to the total area of ​​the bottom detection area; When the coverage area ratio is greater than the second ratio, the surface attachments of the bottom detection area are cleaned and the image of the bottom detection area is re-captured; When the coverage area ratio is less than or equal to the second ratio, the defect features present in the bottom detection area are determined based on the area of ​​the continuous blank area. When the area of ​​a continuous blank region in the bottom detection region image is greater than the first area threshold, it is determined that there is a peeling defect in the bottom detection region. When the area of ​​a continuous blank region in the bottom detection region image is less than the first area threshold, it is determined that there is no peeling defect in the bottom detection region.

9. A rapid defect detection system for bridge piers used by a pier-climbing robot, characterized in that, A rapid defect detection method for bridge piers using a climbing robot, as described in any one of claims 1-8, includes: The camera brightness adjustment module is used to obtain the surface hardness and carbonation depth of the pier column, obtain the concrete strength grade and aggregate type of the pier column based on the surface hardness and carbonation depth, and adjust the brightness and contrast of the camera mounted on the pier climbing robot based on the concrete strength grade and aggregate type. The camera position adjustment module is used to acquire the arithmetic mean deviation of the roughness and the flatness data of the pier surface, and adjust the shooting parameters of the camera based on the arithmetic mean deviation of the roughness, and adjust the position of the camera based on the flatness data. The regional defect detection module is used to divide the pier into detection areas for defect detection, and adjust the defect detection method based on the detection areas. The detection areas include: a top detection area, a middle detection area, and a bottom detection area. The defect result generation module is used to determine the defect level of the bridge pier based on the defect characteristics of each detection area, and generate the bridge pier defect detection results.

Images