A grouting pipe automatic detection method and system based on an intelligent inspection robot

Abstract

This invention relates to the field of intelligent inspection technology for grouting projects, and particularly to an automatic inspection method and system for grouting pipes based on an intelligent inspection robot. The method includes the following steps: collecting inspection environment data using an intelligent inspection robot, and using this data to determine the location of the grouting pipe to be inspected and the inspection movement path; controlling the intelligent inspection robot to move towards the location of the grouting pipe to be inspected along the inspection movement path, and acquiring multi-angle images of the grouting pipe; identifying surface damage to the grouting pipe based on the multi-angle images, and calculating the grouting pipe damage rate; detecting grout leakage based on the grouting pipe damage rate, and recording it as a leakage anomaly; associating the leakage anomaly with the location of the grouting pipe to be inspected, and delineating defect areas in the grouting pipe. This invention, based on intelligent inspection technology for grouting projects, achieves automatic detection of grouting pipe damage and grout leakage, improving the accuracy of grouting pipe defect identification and inspection efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent inspection technology for grouting projects, and in particular to an automatic inspection method and system for grouting pipes based on an intelligent inspection robot. Background Technology

[0002] In current grouting project quality inspection processes, manual inspections and periodic sampling are commonly used to check the condition of grouting pipes. Due to the complex distribution of grouting pipes, the construction environment, the concealment of the pipes, and limitations of manual inspection, problems such as difficulty in detecting surface damage, inaccurate location of grout leakage, and low inspection efficiency are common. Although some technologies use image detection or robotic inspection for auxiliary detection, most focus on identifying single defects and lack a comprehensive analysis method that combines the spatial location of the grouting pipe, surface damage characteristics, and grout leakage status, making it difficult to achieve automated detection and precise location of grouting pipe defects. Summary of the Invention

[0003] Therefore, it is necessary for the present invention to provide an automatic detection method and system for grouting pipes based on an intelligent inspection robot, so as to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, an automatic inspection method for grouting pipes based on an intelligent inspection robot includes the following steps: Step S1: Collect inspection environment data through the intelligent inspection robot, and use the inspection environment data to determine the location of the grouting pipe to be inspected and the inspection movement path; Step S2: Control the intelligent inspection robot to move towards the location of the grouting pipe to be inspected along the inspection movement path and collect multi-angle images of the grouting pipe; identify surface damage of the grouting pipe based on the multi-angle images and calculate the grouting pipe damage rate; Step S3: Detect grout leakage based on the grouting pipe breakage rate and record it as leakage anomaly; associate the leakage anomaly with the location of the grouting pipe to be tested and delineate the defect area of ​​the grouting pipe. Step S4: Control the intelligent inspection robot to re-inspect the defect area of ​​the grouting pipe and determine the abnormal detection results; package the abnormal detection results into an automatic inspection record of the grouting pipe, wherein the automatic inspection record of the grouting pipe includes the location of the grouting pipe defect, the defect type and the detection time.

[0005] Preferably, this specification also provides an automatic grouting pipe detection system based on an intelligent inspection robot, used to execute the automatic grouting pipe detection method based on an intelligent inspection robot as described above. The automatic grouting pipe detection system based on an intelligent inspection robot includes: The inspection environment scanning module is used to collect inspection environment data through the intelligent inspection robot, and to use the inspection environment data to determine the location of the grouting pipe to be inspected and the inspection movement path. The image acquisition module is used to control the intelligent inspection robot to move towards the grouting pipe to be inspected along the inspection path and acquire multi-angle images of the grouting pipe; based on the multi-angle images of the grouting pipe, it identifies surface damage to the grouting pipe and calculates the grouting pipe damage rate; The leakage detection module is used to detect grout leakage based on the grouting pipe damage rate and record it as leakage anomaly; the leakage anomaly is associated with the location of the grouting pipe to be tested, and the defect area of ​​the grouting pipe is divided. The re-inspection module is used to control the intelligent inspection robot to re-inspect the defective area of ​​the grouting pipe and determine the abnormal detection results. The abnormal detection results are packaged into an automatic inspection record of the grouting pipe, which includes the location of the grouting pipe defect, the defect type, and the detection time.

[0006] The beneficial effects of this invention are as follows: (1) By acquiring inspection environment data through intelligent inspection robots, and combining visual positioning data of grouting pipes, spatial distance data and inspection movement paths, automatic identification of grouting pipe positions and inspection path planning are realized in complex construction environments. This transforms the grouting pipe inspection from the traditional manual inspection method to an automated inspection mode based on robot autonomous positioning and movement detection, thereby improving the efficiency and accuracy of grouting pipe inspection.

[0007] (2) By acquiring images of the grouting pipe from multiple angles and combining the characteristics of changes in the brightness of the pipe wall pixels, the intensity of light reflection, and the changes in surface roughness, the peeling area on the surface of the grouting pipe is identified and the damage rate is calculated. This realizes the automatic detection of defects on the surface of the grouting pipe, transforming the traditional damage judgment method that relies on manual observation into a detection mode based on image features and analysis of the physical state of the pipe wall, thereby improving the accuracy of grouting pipe damage identification.

[0008] (3) By extracting grout leakage characteristic data based on the location of the grouting pipe damage, and combining the grout diffusion direction, diffusion range and damage location for correlation analysis, the automatic determination of abnormal grout leakage areas is realized, which transforms the grouting pipe detection from a single surface defect identification to a comprehensive detection mode that combines structural damage and grout migration status, thereby improving the grouting pipe defect location accuracy and engineering quality detection reliability. Attached Figure Description

[0009] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the steps of an automatic inspection method for grouting pipes based on an intelligent inspection robot according to the present invention. Figure 2 This is a schematic diagram of the distance measurement process in this invention; Figure 3This is a schematic diagram of the process for calculating the grouting pipe breakage rate in this invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0010] To achieve the above objectives, please refer to Figures 1 to 3 This invention provides an automatic inspection method for grouting pipes based on an intelligent inspection robot, the method comprising the following steps: Step S1: Collect inspection environment data through the intelligent inspection robot, and use the inspection environment data to determine the location of the grouting pipe to be inspected and the inspection movement path; In one embodiment, an intelligent inspection robot equipped with an industrial camera, laser rangefinder, and inertial measurement unit (IMU) inspects and collects data on the grouting construction area, acquiring images of the inspection environment, spatial distance information, and robot posture information. The industrial camera captures images of the construction area, identifying exposed areas of the grouting pipe through image grayscale changes, edge contours, and pipe continuity features. Based on the identified grouting pipe contour information, the laser rangefinder is controlled to sample distances along the pipe's direction, obtaining multi-point spatial distance data between the robot and the grouting pipe. Spatial coordinate transformation is performed based on these multiple spatial distance data to determine the grouting pipe's position coordinates within the inspection environment. Subsequently, the grouting pipe's extension direction is analyzed based on its position coordinates, and the robot's movement nodes are generated by combining this with the obstacle distribution within the inspection area. An inspection movement path is then planned from the current position to the location of the grouting pipe to be inspected.

[0011] In another embodiment, it is assumed that the intelligent inspection robot collects 200 frames of environmental images in the grouting construction area and identifies 5 exposed areas of the grouting pipe, of which 3 are located on the wall surface and 2 are located on the ground. The robot collects 50 sets of spatial distance points of the grouting pipe through a laser rangefinder sensor, and determines 20 coordinate points of the center position of the grouting pipe after spatial correlation. The robot calculates the extension direction of the pipe based on the center position points of the grouting pipe, and generates 15 inspection path nodes by combining the location data of 10 obstacles in the detection area, thus forming a movement path from the robot's starting position to the grouting pipe detection area.

[0012] Step S2: Control the intelligent inspection robot to move towards the location of the grouting pipe to be inspected along the inspection movement path and collect multi-angle images of the grouting pipe; identify surface damage of the grouting pipe based on the multi-angle images and calculate the grouting pipe damage rate; In one embodiment, an intelligent inspection robot is controlled to move along an inspection path to the location of the grouting pipe to be inspected. A multi-angle image acquisition device installed on the side of the robot performs circumferential imaging of the grouting pipe, acquiring surface images of the grouting pipe from different perspectives. The multi-angle images of the grouting pipe are converted to grayscale, and pixel brightness distribution data within the pipe wall area is extracted. Based on changes in pixel brightness, areas of missing material on the pipe wall surface are identified, and color anomalies are extracted. The changes in light reflection intensity corresponding to the color anomalies are used to analyze changes in the surface roughness of the pipe wall, and the peeling areas on the grouting pipe surface are determined based on the roughness anomalies. The edges of the peeling areas are enhanced, and the closed contours of the peeling areas are extracted. The pixel area of ​​the damaged area is calculated based on the contour range, and the grouting pipe damage rate is calculated in conjunction with the surface area of ​​the grouting pipe.

[0013] In another embodiment, it is assumed that the intelligent inspection robot collects image data of the grouting pipe in four circumferential directions, obtaining a total of 120 images of the grouting pipe from multiple angles; after grayscale processing, 35 abnormal pixel brightness areas are extracted, of which 20 have continuous brightness changes; the light reflection intensity is calculated based on the color abnormal points, resulting in 18 abnormal reflection areas, and 12 areas with surface roughness changes are analyzed; 12 peeling boundary contours are obtained through edge enhancement processing, of which the largest peeling area is 3500 pixels, and the total surface area of ​​the grouting pipe is 50000 pixels. Based on this, the damage rate of the grouting pipe is calculated to be 7%.

[0014] Step S3: Detect grout leakage based on the grouting pipe breakage rate and record it as leakage anomaly; associate the leakage anomaly with the location of the grouting pipe to be tested and delineate the defect area of ​​the grouting pipe. In one embodiment, the damaged areas of the grouting pipe wall are determined based on the pipe breakage rate, and the grout leakage characteristics at the corresponding locations are extracted. Based on the breakage rate data, pipe wall damage areas exceeding a set range are filtered to obtain suspected leakage detection areas. Surface color change data is extracted from these suspected leakage detection areas to identify color shift areas formed after grout seepage. Pixel brightness attenuation and texture changes are analyzed based on the color shift areas to obtain grout adhesion characteristics. Wet areas of the pipe wall are extracted based on the grout adhesion characteristics, and the diffusion direction of the wet areas along the pipe wall surface is analyzed to obtain grout diffusion path data. Spatial matching is performed between the grout diffusion path data and the pipe wall breakage location. When a correspondence exists between the grout diffusion area and the breakage area, an abnormal grout leakage is identified, and the abnormal location is associated with the spatial location of the grouting pipe, forming a grouting pipe defect area.

[0015] In another embodiment, assuming that 8 damaged areas of the grouting pipe are detected in step S2, of which 5 have a damage rate exceeding the preset detection range; extract color change features from the 5 damaged areas to obtain 4 areas of grout color shift; analyze the surface reflection change data based on the color shift areas to identify 3 areas of grout adhesion; extract the wetting diffusion area to obtain grout diffusion ranges of 120mm², 80mm², and 150mm² respectively; after spatially matching the diffusion area with the damaged location, determine that 3 of them have abnormal grout leakage, and divide them into 3 grouting pipe defect areas.

[0016] Step S4: Control the intelligent inspection robot to re-inspect the defect area of ​​the grouting pipe and determine the abnormal detection results; package the abnormal detection results into an automatic inspection record of the grouting pipe, wherein the automatic inspection record of the grouting pipe includes the location of the grouting pipe defect, the defect type and the detection time.

[0017] In one embodiment, an intelligent inspection robot is controlled to move to the defect area of ​​the grouting pipe for re-inspection, and high-definition images of the defect area are acquired through a close-range image acquisition device. The defect area is re-analyzed based on the re-inspection images to extract the defect boundary, grout adhesion state, and defect location coordinates, and the initial inspection results are corrected. Abnormal inspection results are determined based on the re-inspection results, and the data is encapsulated according to the spatial location of the grouting pipe, the defect type, and the inspection time to generate an automatic inspection record of the grouting pipe.

[0018] In another embodiment, it is assumed that the re-inspection robot performs secondary inspections on three defective areas of the grouting pipe, and collects 30 high-definition re-inspection images for each area, obtaining a total of 90 re-inspection data. After re-inspection, it is confirmed that two of the three defective areas are pipe wall peeling defects and one is grout leakage defect. Three sets of corresponding defect spatial location coordinates and three sets of detection time information are recorded, and the defect location, defect type and detection time are associated and stored to form an automatic inspection record of the grouting pipe.

[0019] Preferably, step S1 includes: The image acquisition device on the intelligent inspection robot is used to scan the grouting construction area, identify the exposed area of ​​the grouting pipe, the outline of the pipe body and the position of the interface, and obtain the visual positioning data of the grouting pipe. In one embodiment, an industrial camera mounted on an intelligent inspection robot continuously acquires images of the grouting construction area. The acquired images are then spatially corrected using camera calibration parameters to obtain two-dimensional image data of the grouting construction area. The two-dimensional image data is then converted to grayscale to extract the edge gradient change features between the grouting pipe and the surrounding construction environment. Based on these edge gradient change features, the pipe boundary with a continuous arc structure is identified, and the exposed area of ​​the grouting pipe is extracted. The axial extension direction of the grouting pipe is analyzed based on the spatial continuity of the pipe boundary to obtain the grouting pipe contour data. Based on the grouting pipe contour data, the morphological change features at the pipe connection points are detected to identify the grouting pipe interface area. The exposed area, pipe contour, and interface location are spatially correlated to establish visual positioning data for the grouting pipe.

[0020] In another embodiment, assuming the intelligent inspection robot is equipped with an industrial camera with a resolution of 1920×1080, it acquires 150 consecutive images of the grouting construction area; after edge feature extraction, it identifies 8 exposed areas of grouting pipes, including 5 horizontally arranged grouting pipes and 3 vertically arranged grouting pipes; based on the continuity of the pipe outline, it extracts 320 sets of grouting pipe boundary points and fits 12 pipe center lines; it detects changes in pipe diameter and areas of abrupt grayscale changes at connection positions, and identifies 10 grouting pipe interface positions; it associates the exposed areas, outline boundaries, and interface coordinates to form visual positioning data of the grouting pipes.

[0021] The distance detection device of the intelligent inspection robot, controlled by the visual positioning data of the grouting pipe, measures the distance to obtain spatial distance data; and determines the position of the grouting pipe to be inspected based on the spatial distance data. In one embodiment, based on the pipe center position and interface position in the visual positioning data of the grouting pipe, the laser ranging device mounted on the intelligent inspection robot is controlled to emit ranging signals towards the surface of the grouting pipe, and the signal return time and propagation distance information are collected. The single-point distance between the robot and the surface of the grouting pipe is calculated based on the ranging signal return time; multiple single-point distances are correlated according to the pipe extension direction in the visual positioning data of the grouting pipe to form spatial distance distribution data of the grouting pipe surface. The relative spatial relationship between the robot and the grouting pipe is analyzed based on the spatial distance distribution data to determine the three-dimensional spatial coordinate position of the grouting pipe, and the corresponding position is used as the position of the grouting pipe to be inspected.

[0022] In another embodiment, assuming the intelligent inspection robot determines six locations in the central area of ​​the grouting pipe based on visual positioning data, and controls a laser ranging device to sample distances in each of the six areas, collecting 20 distance points in each area, resulting in a total of 120 sets of single-point distance data; the distance data ranges from 0.8m to 2.5m. Based on the spatial coordinates of the distance points, 30 sets of spatial location points of the grouting pipe are calculated, with the near end of the grouting pipe 0.85m from the robot and the far end 2.35m from the robot; based on the spatial distance distribution, three sets of coordinates for the grouting pipe to be inspected are determined.

[0023] Based on the location of the grouting pipe to be inspected, the extension direction of the grouting pipe and the distribution of surrounding obstacles are analyzed to determine the robot's movement direction and obstacle avoidance position, and to generate an inspection movement path.

[0024] In one embodiment, the positional change relationship between adjacent spatial points is extracted based on the location data of the grouting pipe to be inspected, and the axial extension direction of the grouting pipe is calculated; the movement direction of the robot during the inspection process is determined based on the extension direction of the grouting pipe. The environmental perception device on the intelligent inspection robot collects the location data of surrounding obstacles, and analyzes the interference relationship between the obstacles and the movement direction based on the spatial coordinates of the obstacles; when the obstacle area overlaps with the robot's movement area, the obstacle avoidance position is determined, and the robot's movement nodes are adjusted based on the obstacle avoidance position; the movement direction, obstacle avoidance position, and the location of the grouting pipe to be inspected are connected to generate an inspection movement path.

[0025] In another embodiment, it is assumed that 40 sets of spatial coordinate points of the grouting pipe are obtained based on the location data of the grouting pipe to be detected, and the three main extension directions of the grouting pipe are calculated by the changes of adjacent coordinate points; the inspection area is scanned using an environmental sensing device to identify 15 surrounding obstacles, of which 5 obstacles are within the safe range of the movement path; 8 obstacle avoidance nodes are calculated based on the obstacle positions, and 20 robot movement nodes are replanned; the starting position, obstacle avoidance nodes, and grouting pipe detection position are connected to generate an inspection movement path with a length of approximately 35m.

[0026] Preferably, the distance detection device of the intelligent inspection robot, controlled by the visual positioning data of the grouting pipe, performs distance measurement to obtain spatial distance data including: The center position of the grouting pipe is determined by visual positioning data of the grouting pipe, and the distance detection device of the intelligent inspection robot is controlled to send a detection signal to the center position of the grouting pipe, and the return time and propagation distance of the detection signal are collected. In one embodiment, based on the exposed area, pipe outline, and interface position of the grouting pipe in the visual positioning data, the image coordinates of the grouting pipe are spatially mapped to determine the center position of the grouting pipe in the inspection environment. The intelligent inspection robot is controlled to adjust the detection direction of the distance detection device so that the emission direction of the detection device corresponds to the center position of the grouting pipe. A laser detection signal is emitted towards the center position of the grouting pipe through the distance detection device, and the return time and propagation distance information after the detection signal contacts the surface of the grouting pipe are collected. Based on the return time sequence and propagation distance sequence obtained from multiple detections, abnormal return signals are eliminated to form the detection signal data corresponding to the center region of the grouting pipe.

[0027] In another embodiment, assuming that the center positions of the grouting pipe are determined at 5 locations using visual positioning data, the distance detection device is controlled to emit detection signals to the 5 center positions respectively, and 10 sets of return time data are continuously collected at each position, resulting in a total of 50 sets of detection signal return data; the effective return time range is 8ns to 25ns, corresponding to a propagation distance range of 1.2m to 3.8m; by filtering the return time data, 3 sets of abnormal reflection signals are removed, and 47 sets of effective detection signal data are retained for calculating the distance between the robot and the grouting pipe.

[0028] The distance between the intelligent inspection robot and the surface of the grouting pipe is calculated based on the signal return time and propagation distance information to obtain single-point distance data; multiple single-point distance data are correlated to obtain spatial distance data.

[0029] In one embodiment, the propagation time of the detection signal between the robot and the surface of the grouting pipe is determined based on the signal return time data, and the interval distance between the robot and the surface of the grouting pipe is calculated by combining the propagation speed of the detection signal, thus obtaining single-point distance data. Based on the center position coordinates in the visual positioning data of the grouting pipe, multiple single-point distance data are mapped to corresponding spatial positions; adjacent distance points are sorted according to the axial extension direction of the grouting pipe, and the positional continuity relationship between each distance point is analyzed; single-point distance data with spatial correlation are fused to form spatial distance data of the grouting pipe surface, determining the three-dimensional spatial distribution of the grouting pipe within the inspection area.

[0030] In another embodiment, assuming 47 sets of valid detection signal data are acquired, 47 sets of single-point distance data are calculated based on the return time, where the distance values ​​are distributed in the range of 1.25m to 3.75m; the single-point distance data are correlated according to the extension direction of the grouting pipe, and the 47 distance points are divided into 4 groups of spatially continuous regions, where the first group contains 12 distance points, the second group contains 15 distance points, the third group contains 11 distance points, and the fourth group contains 9 distance points; through spatial coordinate transformation, a total of 47 spatial position points on the surface of the grouting pipe are obtained, and the spatial distance data of the grouting pipe is formed.

[0031] Preferably, step S2, which involves identifying surface damage to the grouting pipe based on multi-angle images and calculating the grouting pipe damage rate, includes: The images of the grouting pipe from multiple angles are converted to grayscale to extract pixel brightness distribution features. The peeling areas on the pipe wall surface are detected based on the pixel brightness distribution features. Edge enhancement processing is performed on the peeling areas on the pipe wall surface to extract the peeling boundary contours. In one embodiment, an intelligent inspection robot is controlled to collect surface images of the grouting pipe from different directions, and the multi-angle images of the grouting pipe are converted to grayscale, converting the color information in the color images into corresponding pixel brightness values. Based on the difference in material reflection between the normal pipe wall area and the peeling area, the pixel brightness distribution within the pipe wall area is statistically analyzed, and the mean pixel brightness, brightness change gradient, and local brightness abrupt change features are extracted. Based on the pixel brightness distribution features, the continuity of the pipe wall surface is analyzed. When a local area shows a brightness abrupt change and a significant difference in brightness from the surrounding pipe wall, it is marked as a suspected peeling area. Valid peeling areas are selected based on the continuous pixel distribution within the suspected peeling area. Subsequently, the edge position of the peeling area is enhanced by enhancing the edge gradient change to highlight the boundary between the peeling area and the intact pipe wall, connecting adjacent edge points and removing isolated noise points to obtain a continuous peeling boundary contour.

[0032] In another embodiment, it is assumed that the intelligent inspection robot collects image data from six circumferential angles of the grouting pipe, obtaining a total of 180 multi-angle images of the grouting pipe. After grayscale processing of the images, a total of 120,000 pixels in the pipe wall area are extracted. Among them, the pixel brightness in the normal area is concentrated in the range of 80 to 150, and the pixel brightness in the abnormal area is distributed in the range of 30 to 70. Based on the brightness change feature detection, 25 suspected peeling areas are obtained. After continuity analysis, 18 effective peeling areas are retained. Edge enhancement processing is performed on the 18 peeling areas, 3,600 edge candidate points are extracted, and 500 isolated abnormal points are removed to form 18 sets of continuous peeling boundary contours.

[0033] The damaged pixel area in a single image is determined by the number of pixels within the peeling boundary contour; the grouting pipe damage rate is calculated based on the damaged pixel area and the multi-angle grouting pipe images.

[0034] In one embodiment, based on the obtained peeling boundary contour, the pixels within each closed contour range are statistically analyzed, and the number of pixels located inside the peeling boundary is taken as the damaged pixel area in a single image. Based on the visible area of ​​the pipe wall corresponding to the multi-angle grouting pipe images, the damaged pixel areas in images from different angles are spatially correlated to avoid repeated calculations of the same damaged area from multiple perspectives. The effective damaged pixel areas in each angle image are accumulated and compared with the pixel area of ​​the overall detection area of ​​the grouting pipe to calculate the grouting pipe damage rate.

[0035] In another embodiment, it is assumed that 18 sets of peeling boundary contours are obtained in step S1, wherein the number of pixels inside the peeling contour in a single image is 320, 450, 680, 520, 760, etc., and a total of 8750 damaged pixel areas are obtained; the total pixel area of ​​the effective detection area corresponding to the multi-angle grouting pipe images is 150,000 pixels; after position association between images from different angles, 1200 pixels of duplicate damaged areas are removed, and the effective damaged pixel area is determined to be 7550 pixels; the grouting pipe damage rate is calculated to be 5.03% based on the effective damaged pixel area and the total detection area.

[0036] Preferably, detecting the peeling area on the pipe wall surface based on pixel brightness distribution characteristics includes: Based on the comparison between the pixel brightness distribution characteristics and the preset pixel brightness threshold, areas of brightness abrupt change are marked; and abnormal color points on the surface of the grouting pipe are identified in the areas of brightness abrupt change. In one embodiment, the brightness variation of each pixel within the grouting pipe wall area is analyzed based on the pixel brightness distribution characteristics obtained after grayscale processing of multi-angle grouting pipe images. The brightness value of each pixel is compared with a preset pixel brightness threshold to determine the magnitude of brightness variation between adjacent pixel areas. When the brightness difference within a local area exceeds a preset range, the corresponding area is marked as a brightness abrupt change area. The color distribution characteristics within the brightness abrupt change area are extracted, and the color information of each pixel is converted into a color space. The degree of color shift between abnormal areas and normal pipe wall areas is analyzed. Based on the color shift, color anomalies caused by pipe wall material peeling, grout adhesion, or surface oxidation are identified, and the location coordinates and corresponding brightness information of the color anomalies are recorded to form grouting pipe surface color anomaly data.

[0037] In another embodiment, it is assumed that the acquired multi-angle grouting pipe images are processed to obtain a total of 80,000 effective detection area pixels of the pipe wall; the pixel brightness threshold range is set to 60-180 according to the brightness distribution of the normal pipe wall area. When more than 20 consecutive pixels in a local area are detected with a brightness lower than 60 or higher than 180, they are marked as brightness change areas. A total of 35 brightness change areas are detected; color feature analysis is performed on the 35 areas, and 18 areas with obvious color shifts are extracted. The number of color anomalies are 120, 85, and 210 respectively, and a total of 1,500 color anomalies are obtained, and their corresponding spatial locations are recorded.

[0038] The light reflection intensity is calculated using the color anomalies on the surface of the grouting pipe, and the surface roughness of the pipe wall is determined based on the light reflection intensity; the peeling area on the pipe wall surface is identified based on the surface roughness.

[0039] In one embodiment, the reflection changes of abnormal points on the grouting pipe surface are analyzed based on the pixel brightness values ​​and color change information corresponding to the abnormal points. The brightness response values ​​under illumination from different directions within the abnormal point area are calculated, and these brightness response values ​​are used as the light reflection intensity on the grouting pipe surface. The change in surface smoothness of the pipe wall is analyzed based on the difference in reflection intensity between the normal and abnormal areas. When the grouting pipe surface peels off, the loss of pipe wall material causes changes in the surface microstructure, resulting in a dispersion of light reflection direction, manifested as a decrease in local reflection intensity or an increase in fluctuation. The surface roughness distribution of the pipe wall is determined based on the changes in light reflection intensity. The peeling areas on the pipe wall surface are identified based on the continuous range and spatial distribution of the roughness abnormal areas.

[0040] In another embodiment, it is assumed that 1500 color anomaly points are obtained, of which 12 are effective anomaly regions distributed on the surface of the grouting pipe. The light reflection intensity of each region is calculated based on the brightness response value of the anomaly points, where the reflection intensity range of the normal pipe wall region is 0.75 to 0.90, and the reflection intensity range of the anomaly region is 0.35 to 0.60. Based on the change in reflection intensity, 10 regions of pipe wall roughness change are calculated, of which 8 regions have roughness exceeding the normal range. Combining the continuous distribution of roughness regions, 8 peeling regions on the surface of the grouting pipe are determined, with peeling region areas of 120 mm², 85 mm², 210 mm², etc.

[0041] Preferably, edge enhancement processing is performed on the peeling area of ​​the pipe wall surface, and the peeling boundary contour is extracted, including: Candidate points of the spalling edge are extracted from the spalling area on the pipe wall surface; edge enhancement processing is performed based on the spatial continuous distribution relationship of the candidate points of the spalling edge, adjacent edge points are connected and isolated outliers are removed to obtain continuous spalling edge lines; In one embodiment, edge gradient detection is performed on the image of the peeled area on the surface of the grouting pipe. The boundary pixels between the peeled area and the intact pipe wall are extracted based on the direction and magnitude of pixel brightness changes. Pixels with significant gradient changes are selected as candidate peeling edges. A spatially continuous distribution relationship is established based on the adjacency relationship of each candidate edge in the image coordinates. The connection distance and direction change between adjacent candidate points are calculated, and candidate points that satisfy the continuous change relationship are sequentially connected. For edge locations with large connection distances or abrupt changes in direction, edge compensation is performed based on the positional relationship of adjacent points. Simultaneously, isolated abnormal points without continuous connection relationships are removed to enhance the integrity of the peeling edges, resulting in continuous peeling edge lines.

[0042] In another embodiment, assuming that an image of the surface of a grouting pipe is processed, 500 candidate points for spalling edges are extracted. The spatial distribution of these candidate points is analyzed based on 8-neighborhood connectivity. 420 candidate points satisfy the continuous connectivity condition, while 80 candidate points are marked as outliers and removed due to insufficient connections with surrounding areas. Subsequently, connection compensation is performed on the intervals between continuous edge points, resulting in four continuous spalling edge lines. The longest edge line contains 260 continuous edge points, and the shortest edge line contains 45 continuous edge points.

[0043] The boundary range of the spalling area is determined based on the continuous spalling edge line, and the contour is reconstructed within the boundary range of the spalling area to obtain the spalling boundary contour.

[0044] In one embodiment, the outer range of the peeling region is determined based on the start and end positions of the continuous peeling edge lines and the direction of edge extension. Pixels located within the area enclosed by the continuous peeling edge lines are then filled to form an initial peeling region. The contours are sorted according to the connection relationships between the continuous peeling edge lines, and multiple continuous edge segments are spliced ​​together in spatial order to construct a closed boundary. For gaps at the boundary connections, the contours are filled in according to the directional change trends of adjacent edge points, ensuring that the reconstructed contours match the actual peeling region boundary, thus obtaining a complete peeling boundary contour.

[0045] In another embodiment, assuming that after continuous peeling edge line analysis, three peeling regions are identified on the surface of the grouting pipe, wherein the first peeling region consists of 180 edge points, the second peeling region consists of 95 edge points, and the third peeling region consists of 60 edge points. Contour reconstruction is performed based on the spatial arrangement of the edge points in each region, and adjacent edge points are connected in a clockwise direction to form three closed contours, with corresponding contour areas of 2100 pixels, 860 pixels, and 430 pixels, respectively.

[0046] Preferably, in step S3, grout leakage detection is performed based on the grouting pipe breakage rate, and the leakage anomalies are recorded as follows: The location of pipe wall damage is determined based on the pipe damage rate; surface humidity change characteristics, color change characteristics, and grout adhesion characteristics are extracted based on the location of pipe wall damage to obtain grout leakage characteristic data; the diffusion state of grout at the location of pipe wall damage is analyzed based on the grout leakage characteristic data. In one embodiment, based on the calculated grouting pipe breakage rate, the pipe wall areas with breakage rates exceeding a preset detection range are subjected to reverse mapping, converting the corresponding image coordinates into spatial coordinates to determine the pipe wall breakage location. Subsequently, a local detection area is established around the breakage location, and features are extracted from the pipe wall surface within this local area. A humidity detection device is used to collect surface humidity information of the breakage area and surrounding areas, analyzing humidity gradient changes to obtain surface humidity change characteristics. Simultaneously, color change features are extracted based on pixel color changes in multi-angle images, and grout adhesion areas on the pipe wall surface are identified through grout reflection features and adhesion texture features, forming grout leakage feature data that includes humidity changes, color changes, and grout adhesion status. Based on the humidity change trend, color anomaly range, and grout adhesion continuity at different locations, the direction and range of grout diffusion from the breakage location to the surrounding area are analyzed to determine the grout diffusion state at the pipe wall breakage location.

[0047] In another embodiment, it is assumed that three pipe wall damage locations are determined based on the grouting pipe damage rate detection results, located at pipe segment coordinates (120mm, 350mm), (260mm, 420mm), and (510mm, 380mm), respectively. Detection data is collected within a 50mm radius around each damage location. In the first damage area, surface humidity increases from 42% to 68%, and the grayscale value decreases from 135 to 96, with a detected grout adhesion area of ​​320mm². In the second damage area, surface humidity increases from 38% to 55%, and the grayscale value decreases from 128 to 104, with a grout adhesion area of ​​180mm². In the third damage area, surface humidity increases from 40% to 72%, and the grayscale value decreases from 140 to 90, with a grout adhesion area of ​​410mm². Based on the spatial overlap of the humidity increase area, color change area, and grout adhesion area, it is determined that the first and third damage locations show a significant grout diffusion trend, while the diffusion range of the second damage location is smaller.

[0048] The diffusion range of the slurry is determined based on the diffusion state; the location of the slurry diffusion range is matched with the location of the pipe wall damage to identify slurry leakage and mark it as leakage anomaly.

[0049] In one embodiment, based on the analysis results of slurry diffusion status, the humidity change area, color abnormality area, and slurry adhesion area are spatially fused to determine the slurry diffusion boundary. The slurry diffusion range is obtained by extracting the edge points of the diffusion area and connecting them continuously. The center position, boundary range, and diffusion direction of the slurry diffusion range are spatially matched with the determined pipe wall damage location. When the slurry diffusion area covers the damage location or extends outward along the damage location, it is determined that there is slurry leakage at that location, and the corresponding pipe wall area is marked as a leakage abnormality area. Simultaneously, the location coordinates, diffusion range, and associated damage location of the leakage abnormality area are recorded to form the grouting pipe leakage detection result.

[0050] In another embodiment, it is assumed that three slurry diffusion regions are obtained through slurry diffusion state analysis. The first diffusion region has a boundary range of 45mm × 30mm, and its center position deviates from the first damage location by 2mm. The second diffusion region has a boundary range of 20mm × 15mm, and its center position deviates from the second damage location by 18mm. The third diffusion region has a boundary range of 60mm × 35mm, and its center position deviates from the third damage location by 3mm. According to the spatial location matching results, the first and third diffusion regions both cover the corresponding pipe wall damage locations, and the diffusion direction is consistent with the outward extension direction of the damage location. Therefore, they are determined to be slurry leakage anomalies. The second diffusion region is far from the damage location and does not form a continuous diffusion region. Therefore, it is determined to be a non-leakage region. Two leakage anomaly locations are marked, and corresponding anomaly detection records are generated.

[0051] Preferably, based on the location of pipe wall damage, surface humidity change characteristics, color change characteristics, and slurry adhesion characteristics are extracted to obtain slurry leakage characteristic data, including: Extract color change features from pipe wall damage location data; identify color shift areas formed after slurry seepage based on color change features; analyze brightness attenuation and texture changes within the color shift areas based on color shift areas, and record surface reflection changes caused by slurry coverage. In one embodiment, based on the determined pipe wall damage location data, the multi-angle grouting pipe image corresponding to the damage location is used for region localization. Pixel color information is extracted within the localized area, and the difference in color distribution before and after the damage is used as a color change feature. Based on the color change feature, the pixels on the pipe wall surface are classified, and areas with significant changes in color saturation, grayscale value, and hue are marked as color shift areas after grout seepage. Brightness gradient analysis is performed on the color shift areas to calculate the brightness attenuation at different locations within the area. Simultaneously, texture change information is extracted. By comparing the texture continuity, edge sharpness, and surface roughness changes between the grout-covered and uncovered areas, the surface reflection changes caused by grout adhesion are determined, and the surface reflection change data is recorded.

[0052] In another embodiment, assuming three areas to be analyzed are determined based on the location data of pipe wall damage, color change features are extracted from each area. Specifically, in the first damaged area, the average grayscale value decreased from 142 to 96, while the color saturation increased from 0.35 to 0.62; in the second damaged area, the average grayscale value decreased from 136 to 118, while the color saturation increased from 0.38 to 0.49; and in the third damaged area, the average grayscale value decreased from 145 to 88, while the color saturation increased from 0.32 to 0.68. Based on the color shift range, three slurry seepage areas are identified. Brightness attenuation and texture change analysis are performed on each area. The first and third areas show significant brightness reduction and texture blurring, with corresponding reflection intensity changes of 0.71 and 0.65, respectively. The reflection intensity change in the second area is 0.86.

[0053] Wet areas on the pipe wall surface are extracted based on surface reflection change data to determine surface humidity change characteristics; slurry adhesion characteristics are determined based on surface reflection change data; surface humidity change characteristics, color change characteristics, and slurry adhesion characteristics are used as slurry leakage characteristic data.

[0054] In one embodiment, the differences in light reflection intensity in different regions are analyzed based on surface reflection change data, and regions with decreased and continuously distributed reflection intensity are identified as wetting areas on the pipe wall surface. The surface humidity change characteristics caused by slurry seepage are extracted based on the area change, boundary diffusion direction, and reflection intensity change trend of the wetting area. Simultaneously, the adhesion location, adhesion range, and adhesion degree of slurry on the pipe wall surface are determined based on the color residue range, texture coverage, and reflection characteristic changes within the surface reflection change area, forming slurry adhesion characteristics. The surface humidity change characteristics, color change characteristics, and slurry adhesion characteristics are fused to generate slurry leakage characteristic data for slurry leakage analysis.

[0055] In another embodiment, assuming feature fusion analysis is performed on the first, second, and third pipe wall damage areas, the detected area of ​​the first area is a wetted area of ​​420 mm², a color-shifted area of ​​380 mm², and a slurry-attached area of ​​350 mm²; the area of ​​the second area is a wetted area of ​​160 mm², a color-shifted area of ​​140 mm², and a slurry-attached area of ​​120 mm²; and the area of ​​the third area is a wetted area of ​​560 mm², a color-shifted area of ​​510 mm², and a slurry-attached area of ​​470 mm². Based on the changes in reflection intensity, the diffusion range of the wetted area, and the continuity of slurry adhesion in each area, corresponding slurry leakage characteristic data are generated. The first and third areas show obvious humidity diffusion and slurry coverage characteristics, while the second area only shows local color changes and weak adhesion characteristics.

[0056] Preferably, analyzing the diffusion state of the slurry at the location of the pipe wall rupture based on slurry leakage characteristic data includes: Identify the diffusion direction based on grout leakage characteristic data, and determine the grout diffusion path along the diffusion direction; analyze the continuity of grout coverage on the pipe wall surface based on the grout diffusion path, and determine the degree of seepage diffusion based on the coverage continuity; determine the diffusion state of grout at the pipe wall damage location by combining the degree of diffusion, coverage continuity, and diffusion direction.

[0057] In one embodiment, spatial correlation analysis is performed on the leakage area around the pipe wall damage location based on the surface humidity change characteristics, color shift area distribution, and slurry adhesion strength in the slurry leakage feature data. Based on the boundaries of the wetted areas, color change gradients, and the direction of slurry adhesion thickness change at different locations, a slurry diffusion direction determination rule is established. The damage point is used as the initial diffusion source point, and the main diffusion direction of the slurry is determined along the direction of increasing surface humidity gradient. Subsequently, continuous leakage areas are extracted along the determined diffusion direction, and adjacent color shift points and wetted area boundary points are connected to form a slurry diffusion path. The continuous pixel areas on the diffusion path are analyzed, and the overlap, boundary continuity, and slurry adhesion changes between adjacent detection areas are calculated. When adjacent areas remain continuously connected without obvious breaks, it is determined to be a continuous coverage state; when local breaks or rapid attenuation of adhesion strength occur, it is determined to be a discontinuous coverage state. Based on the length of the continuous coverage area, the diffusion path range, and the slurry adhesion change trend, the degree of slurry seepage diffusion is divided, and slurry diffusion state data is generated by combining the diffusion direction, coverage continuity, and diffusion degree.

[0058] In another embodiment, assuming that grout leakage characteristic data of a damaged area of ​​a grouting pipe is obtained through multi-angle image acquisition, where the coordinates of the damaged center point are (320, 240), 500 pixels of the surrounding wet area are detected, and the color offset area covers an area of ​​3200 pixels. Based on the color change gradient, the main diffusion direction of the grout is calculated to shift upwards and to the right from the damaged point, with a diffusion angle of approximately 35°. The length of the continuous diffusion path along this direction is extracted to be 85 mm. Analyzing the coverage continuity along the diffusion path, the path is divided into 10 detection segments, of which 8 segments have continuous grout coverage, and 2 segments have adhesion interruptions, resulting in a coverage continuity rate of 80%. Simultaneously, the highest grout adhesion strength is detected near the damaged center, gradually decreasing towards the diffusion end, determining that this area is at a moderate diffusion level. Based on the diffusion direction, 80% coverage continuity, and moderate diffusion level, the pipe wall damage location is marked as a directional continuous diffusion state.

[0059] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0060] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. An automatic inspection method for grouting pipes based on an intelligent inspection robot, characterized in that, Includes the following steps: Step S1: Collect inspection environment data through the intelligent inspection robot, and use the inspection environment data to determine the location of the grouting pipe to be inspected and the inspection movement path; Step S2: Control the intelligent inspection robot to move towards the location of the grouting pipe to be inspected along the inspection movement path and collect multi-angle images of the grouting pipe; identify surface damage of the grouting pipe based on the multi-angle images and calculate the grouting pipe damage rate; Step S3: Detect grout leakage based on the grouting pipe breakage rate and record it as leakage anomaly; associate the leakage anomaly with the location of the grouting pipe to be tested and delineate the defect area of ​​the grouting pipe; Step S4: Control the intelligent inspection robot to re-inspect the defect area of ​​the grouting pipe and determine the abnormal detection results; package the abnormal detection results into an automatic inspection record of the grouting pipe, wherein the automatic inspection record of the grouting pipe includes the location of the grouting pipe defect, the defect type and the detection time.

2. The automatic detection method for grouting pipes based on an intelligent inspection robot according to claim 1, characterized in that, Step S1 includes: The image acquisition device on the intelligent inspection robot is used to scan the grouting construction area, identify the exposed area of ​​the grouting pipe, the outline of the pipe body and the position of the interface, and obtain the visual positioning data of the grouting pipe. The distance detection device of the intelligent inspection robot, controlled by the visual positioning data of the grouting pipe, measures the distance to obtain spatial distance data; and determines the position of the grouting pipe to be inspected based on the spatial distance data. Based on the location of the grouting pipe to be inspected, the extension direction of the grouting pipe and the distribution of surrounding obstacles are analyzed to determine the robot's movement direction and obstacle avoidance position, and to generate an inspection movement path.

3. The automatic detection method for grouting pipes based on an intelligent inspection robot according to claim 2, characterized in that, The distance detection device of the intelligent inspection robot, controlled by the visual positioning data of the grouting pipe, measures the distance and obtains spatial distance data, including: The center position of the grouting pipe is determined by visual positioning data of the grouting pipe, and the distance detection device of the intelligent inspection robot is controlled to send a detection signal to the center position of the grouting pipe, and the return time and propagation distance of the detection signal are collected. The distance between the intelligent inspection robot and the surface of the grouting pipe is calculated based on the signal return time and propagation distance information to obtain single-point distance data; multiple single-point distance data are correlated to obtain spatial distance data.

4. The automatic detection method for grouting pipes based on an intelligent inspection robot according to claim 1, characterized in that, Step S2 involves identifying surface damage to the grouting pipe based on multi-angle images and calculating the grouting pipe damage rate, including: The images of the grouting pipe from multiple angles are converted to grayscale to extract pixel brightness distribution features. The peeling areas on the pipe wall surface are detected based on the pixel brightness distribution features. Edge enhancement processing is performed on the peeling areas on the pipe wall surface to extract the peeling boundary contours. The damaged pixel area in a single image is determined by the number of pixels within the peeling boundary contour; the grouting pipe damage rate is calculated based on the damaged pixel area and the multi-angle grouting pipe images.

5. The automatic detection method for grouting pipes based on an intelligent inspection robot according to claim 4, characterized in that, Based on pixel brightness distribution characteristics, the detection of peeling areas on the pipe wall surface includes: Based on the comparison between the pixel brightness distribution characteristics and the preset pixel brightness threshold, areas of brightness abrupt change are marked; and abnormal color points on the surface of the grouting pipe are identified in the areas of brightness abrupt change. The light reflection intensity is calculated using the color anomalies on the surface of the grouting pipe, and the surface roughness of the pipe wall is determined based on the light reflection intensity; the peeling area on the pipe wall surface is identified based on the surface roughness.

6. The automatic detection method for grouting pipes based on an intelligent inspection robot according to claim 4, characterized in that, Edge enhancement processing is performed on the spalled area of ​​the pipe wall surface, and the spalled boundary contour is extracted, including: Candidate points of the spalling edge are extracted from the spalling area on the pipe wall surface; edge enhancement processing is performed based on the spatial continuous distribution relationship of the candidate points of the spalling edge, adjacent edge points are connected and isolated outliers are removed to obtain continuous spalling edge lines; The boundary range of the spalling area is determined based on the continuous spalling edge line, and the contour is reconstructed within the boundary range of the spalling area to obtain the spalling boundary contour.

7. The automatic detection method for grouting pipes based on an intelligent inspection robot according to claim 1, characterized in that, In step S3, grout leakage is detected based on the grouting pipe breakage rate, and the leakage anomalies are recorded as follows: The location of pipe wall damage is determined based on the pipe damage rate; surface humidity change characteristics, color change characteristics, and grout adhesion characteristics are extracted based on the location of pipe wall damage to obtain grout leakage characteristic data; the diffusion state of grout at the location of pipe wall damage is analyzed based on the grout leakage characteristic data. The diffusion range of the slurry is determined based on the diffusion state; the location of the slurry diffusion range is matched with the location of the pipe wall damage to identify slurry leakage and mark it as leakage anomaly.

8. The automatic detection method for grouting pipes based on an intelligent inspection robot according to claim 7, characterized in that, Based on the location of pipe wall damage, surface humidity change characteristics, color change characteristics, and slurry adhesion characteristics were extracted to obtain slurry leakage characteristic data, including: Extract color change features from pipe wall damage location data; identify color shift areas formed after slurry seepage based on color change features; analyze brightness attenuation and texture changes within the color shift areas based on color shift areas, and record surface reflection changes caused by slurry coverage. Wet areas on the pipe wall surface are extracted based on surface reflection change data to determine surface humidity change characteristics; slurry adhesion characteristics are determined based on surface reflection change data; surface humidity change characteristics, color change characteristics, and slurry adhesion characteristics are used as slurry leakage characteristic data.

9. The automatic detection method for grouting pipes based on an intelligent inspection robot according to claim 7, characterized in that, Analysis of grout diffusion characteristics at the pipe wall rupture site based on grout leakage data includes: Identify the diffusion direction based on grout leakage characteristic data, and determine the grout diffusion path along the diffusion direction; analyze the continuity of grout coverage on the pipe wall surface based on the grout diffusion path, and determine the degree of seepage diffusion based on the coverage continuity; determine the diffusion state of grout at the pipe wall damage location by combining the degree of diffusion, coverage continuity, and diffusion direction.

10. An automatic inspection system for grouting pipes based on an intelligent inspection robot, characterized in that, For performing the automatic grouting pipe detection method based on an intelligent inspection robot as described in claim 1, the automatic grouting pipe detection system based on the intelligent inspection robot includes: The inspection environment scanning module is used to collect inspection environment data through the intelligent inspection robot, and to use the inspection environment data to determine the location of the grouting pipe to be inspected and the inspection movement path. The image acquisition module is used to control the intelligent inspection robot to move towards the grouting pipe to be inspected along the inspection path and acquire multi-angle images of the grouting pipe; based on the multi-angle images of the grouting pipe, it identifies surface damage to the grouting pipe and calculates the grouting pipe damage rate; The leakage detection module is used to detect grout leakage based on the grouting pipe damage rate and record it as leakage anomaly; the leakage anomaly is associated with the location of the grouting pipe to be tested, and the defect area of ​​the grouting pipe is divided. The re-inspection module is used to control the intelligent inspection robot to re-inspect the defective area of ​​the grouting pipe and determine the abnormal detection results. The abnormal detection results are packaged into an automatic inspection record of the grouting pipe, which includes the location of the grouting pipe defect, the defect type, and the detection time.

Images