Municipal road construction defect inspection system and method based on visual inspection

Abstract

The invention relates to the related technical field of visual inspection, in particular to a municipal road construction defect inspection system and method based on visual inspection, and the method comprises the steps: a scanning module places inspection monitoring equipment at a specific position to scan a target area; the similarity comparison module compares the three-dimensional point cloud data; the association mapping module obtains multispectral texture features; the parameter searching module finds construction parameters; and the defect reminding module integrates multiple data to remind construction defects. The technical problems that hidden defects such as non-dense grouting and material layering are prone to being omitted, a large-area construction area cannot be comprehensively inspected in time, and the road construction progress is affected are solved, the difference between a target construction area and a standard model is analyzed through multi-scale similarity comparison, the structural defects are more comprehensively detected, and the construction efficiency is improved. And the visual detection data and the construction process parameters are subjected to fusion analysis, the construction quality is comprehensively evaluated from multiple angles, and the technical effects of determining the specific position of the defect in the space and carrying out defect positioning reminding are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field related to visual detection, and in particular to a municipal road construction defect inspection system and method based on visual detection. Background Art

[0002] The quality of municipal road construction directly impacts the safety and service life of urban infrastructure. With the increasing complexity of road construction processes (e.g., the increase in hidden projects like anchor grouting), the need for refined control of construction quality is becoming increasingly prominent. Existing road construction inspections primarily rely on sensors (e.g., laser ranging and visible light cameras), resulting in a high rate of missed detection of hidden defects. For example, problems such as internal voids and uneven slurry penetration cannot be detected through surface observation, resulting in a high rate of misjudgment of defects and delayed early warnings.

[0003] In summary, the existing technology has hidden defects such as loose grouting and material stratification that are easily missed, making it impossible to conduct comprehensive inspections of large construction areas in a timely manner, thus affecting the progress of road construction. Summary of the Invention

[0004] This application provides a municipal road construction defect inspection system and method based on visual detection, aiming to solve the technical problems in the existing technology that hidden defects such as loose grouting and material stratification are easily missed, making it impossible to conduct comprehensive inspections of large construction areas in a timely manner, thus affecting the progress of road construction.

[0005] In view of the above problems, the technical solution to implement this application is: On the one hand, the present application provides a municipal road construction defect inspection system based on visual detection, wherein the system includes: a scanning module for setting a first position coordinate of an inspection and monitoring device at an associated mapping of an anchor maintenance point; using the inspection and monitoring device to scan a target construction area of ​​a municipal road; a similarity comparison module for performing a multi-scale similarity comparison between the three-dimensional point cloud data scanned by the inspection and monitoring device and the standard three-dimensional point cloud data stored in a historical qualified sample; an association mapping module for setting a second position coordinate of a multi-spectral imaging device at an associated mapping of an anchor maintenance point; using the multi-spectral imaging device to synchronously obtain a first texture feature corresponding to a visible light band and a second texture feature corresponding to a near-infrared band of the target construction area; a parameter search module for searching for grouting pump operating parameters of the anchor maintenance point during a primary grouting construction process; searching for grouting pump operating parameters of the anchor maintenance point during a secondary grouting construction process; and a defect reminder module for providing a construction defect reminder based on the multi-scale similarity comparison information, the first texture feature corresponding to the visible light band, the second texture feature corresponding to the near-infrared band, combined with the grouting pump operating parameters and the grouting pump operating parameters.

[0006] Preferably, the three-dimensional point cloud data and the standard three-dimensional point cloud data are voxel-gridified to generate multiple resolution voxel grids; the point cloud feature descriptors under each resolution voxel grid in the multiple resolution voxel grids are determined, and the point cloud feature descriptors include at least one of normal direction, curvature, and local point density; and multi-scale similarity comparison is performed at different scales using the point cloud feature descriptors.

[0007] Preferably, the multi-scale similarity comparison index includes any one or more of point pair distance, normal vector consistency, and feature point distribution density.

[0008] Preferably, an RGB image of the target construction area is captured, and color features, edge features and texture frequency features in the RGB image are extracted as first texture features; a near-infrared image of the target construction area is captured, and moisture content features, material density features and hidden defect features in the near-infrared image are extracted as second texture features.

[0009] Preferably, a construction parameter state space is constructed through the grouting pump operating parameters and the grouting pump operating parameters; defect fuzzy probability is derived in the construction parameter state space to obtain the construction defect type and the defect fuzzy probability value; based on the construction defect type and the defect fuzzy probability value, it is determined whether to trigger the construction defect reminder.

[0010] Preferably, the grouting pump operating parameters corresponding to the first grouting construction of the anchor maintenance point, including grouting time, grouting pressure, grouting flow, and slurry ratio, are determined; and the grouting pump operating parameters corresponding to the second grouting construction of the anchor maintenance point, including grouting time, grouting pressure peak, pressure stabilization time, and grouting volume, are determined.

[0011] Preferably, if the construction defect reminder is triggered, the defect range is preliminarily defined through multi-scale similarity comparison information; the boundary area of ​​the preliminarily defined defect range is cropped according to the first texture feature corresponding to the visible light band; and the boundary area of ​​the preliminarily defined defect range is cropped according to the second texture feature corresponding to the near-infrared band.

[0012] Preferably, superpixel segmentation is performed on the RGB image within the preliminarily defined defect range, and the image is divided into multiple superpixel blocks with similar color features; the color histogram features, local binary pattern features and grayscale co-occurrence matrix features of multiple superpixel blocks are extracted to construct the feature vectors of the superpixel blocks; and the boundary area of ​​the preliminarily defined defect range is cropped through the feature vectors of the superpixel blocks and the anisotropic features of the edges of the defective areas.

[0013] Preferably, a similarity threshold τ is defined based on the cosine similarity between the feature vector of the superpixel block and the feature vector of the defect area; multiple superpixel blocks are traversed, and superpixel blocks with cosine similarity less than the similarity threshold τ are removed from the preliminary defined defect range; the edge of the cropped defect area is extracted; the corner points and mutation points on the edge of the cropped defect area are identified, the normal vector field of the boundary pixels is generated, the angle distribution between the normal vector and the local texture direction is determined, and the anisotropic characteristics of the edge of the defect area are evaluated.

[0014] On the other hand, the present application provides a municipal road construction defect inspection method based on visual detection, wherein the method includes: an inspection and monitoring device is set at a first position coordinate associated with an anchor maintenance point; the inspection and monitoring device is used to scan a target construction area of ​​a municipal road; a multi-scale similarity comparison is performed on the three-dimensional point cloud data scanned by the inspection and monitoring device and the standard three-dimensional point cloud data stored in a historical qualified sample; a multi-spectral imaging device is set at a second position coordinate associated with an anchor maintenance point; the multi-spectral imaging device is used to synchronously obtain a first texture feature corresponding to a visible light band and a second texture feature corresponding to a near-infrared band of the target construction area; the grouting pump operating parameters of the anchor maintenance point during the first grouting construction process are searched; the grouting pump operating parameters of the anchor maintenance point during the second grouting construction process are searched; and construction defect reminders are issued based on the multi-scale similarity comparison information, the first texture feature corresponding to the visible light band, the second texture feature corresponding to the near-infrared band, combined with the grouting pump operating parameters and the grouting pump operating parameters.

[0015] In summary, one or more technical solutions provided in this application realize the multi-scale similarity comparison analysis of the differences between the target construction area and the standard model, more comprehensively detect structural defects, and integrate visual inspection data with construction process parameters for analysis, comprehensively evaluate construction quality from multiple angles, and determine the specific location of defects in space to provide defect location reminders. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A structural diagram of a municipal road construction defect inspection system based on visual inspection is provided for this application; Figure 2 A flow chart of a municipal road construction defect inspection method based on visual inspection is provided for this application.

[0017] Description of reference numerals: scanning module M100, similarity comparison module M200, association mapping module M300, parameter search module M400, defect reminder module M500. DETAILED DESCRIPTION

[0018] Example 1 The present application is described in detail below with reference to the accompanying drawings. Figure 1 As shown, the present application provides a municipal road construction defect inspection system based on visual detection, wherein the system includes: The scanning module M100 is used to set the patrol monitoring equipment at the first position coordinate associated with the anchor maintenance point; use the patrol monitoring equipment to scan the target construction area of ​​the municipal road; the similarity comparison module M200 is used to perform multi-scale similarity comparison between the three-dimensional point cloud data obtained by the patrol monitoring equipment and the standard three-dimensional point cloud data stored in the historical qualified samples.

[0019] Specifically, the anchor maintenance point is a pre-set key monitoring position, and the first position coordinate associated with it is the specific installation location of the inspection and monitoring equipment; the inspection and monitoring equipment obtains the three-dimensional point cloud data of the target construction area through three-dimensional scanning technology. These data can accurately reflect the geometric shape, surface characteristics and other information of the construction area; the standard three-dimensional point cloud data stored in historical qualified samples refers to the three-dimensional point cloud data corresponding to the construction area that has been verified to be qualified in previous construction, which contains characteristic information such as construction shape and structure that meets the standards; multi-scale similarity comparison refers to the comparison of the three-dimensional point cloud data of the target construction area with the standard three-dimensional point cloud data at different resolution scales to identify differences and potential defects.

[0020] Execution steps: First, based on the first position coordinates of the anchor maintenance point association mapping, the inspection and monitoring equipment is accurately installed near the target construction area of ​​the municipal road; then, the equipment begins to scan the target construction area to obtain detailed 3D point cloud data; then, the scanned 3D point cloud data is compared with the standard 3D point cloud data of the stored historical qualified samples for multi-scale similarity.

[0021] Through voxel gridding and hierarchical processing, the data is divided into voxel grids of different resolutions, and the point cloud feature descriptors at each resolution, such as normal direction and curvature, are determined. Similarity comparison is performed at multiple scales. By performing multi-scale similarity comparison on the three-dimensional point cloud data of the target construction area, it is found that compared with the standard data, there are smaller-scale depressions in the target area. The depression is difficult to detect in low-resolution comparison, but is accurately identified in high-resolution comparison, thereby effectively improving the recognition rate of minor defects and providing an accurate basis for subsequent construction defect reminders.

[0022] The association mapping module M300 is used to set the multispectral imaging device at the second position coordinate of the association mapping of the anchor maintenance point; use the multispectral imaging device to synchronously obtain the first texture features corresponding to the visible light band and the second texture features corresponding to the near-infrared band of the target construction area; the parameter search module M400 is used to search the grouting pump operating parameters of the anchor maintenance point during the first grouting construction process; search the grouting pump operating parameters of the anchor maintenance point during the second grouting construction process; the defect reminder module M500 is used to provide construction defect reminders based on multi-scale similarity comparison information, the first texture features corresponding to the visible light band, the second texture features corresponding to the near-infrared band, combined with the grouting pump operating parameters and the grouting pump operating parameters.

[0023] Specifically, anchor maintenance points are key locations for monitoring and evaluation in municipal road construction, and the second position coordinates of their associated mappings are used to determine the installation location of the multispectral imaging equipment; the multispectral imaging equipment can simultaneously capture images in the visible light and near-infrared bands to extract the texture features of the target construction area, among which the visible light band is used to capture explicit defects such as surface cracks; the near-infrared band is used to detect hidden defects such as moisture penetration and internal voids; the operating parameters of the grouting pump and grouting pump include grouting time, grouting pressure, grouting flow, etc., which reflect the quality and status of grouting and grouting during the construction process.

[0024] Execution steps: First, set the multispectral imaging device at the second position coordinate associated with the anchor maintenance point to ensure that the device can effectively cover the target construction area; then, the device synchronously obtains the first texture features corresponding to the visible light band and the second texture features corresponding to the near-infrared band of the target construction area. The visible light images and near-infrared images obtained by the multispectral imaging device are used to extract color features, edge features, and moisture content features, etc. These features can comprehensively reflect the surface and internal conditions of the construction area and improve the detection rate of hidden defects.

[0025] At the same time, the system searches for the operating parameters of the grouting pump and grouting pump during the first grouting construction process and the second grouting construction process at the anchor maintenance point. These parameters are used to deduce the fuzzy probability of defects through the construction parameter state space, and combined with multi-scale similarity comparison information and texture features, to conduct a comprehensive assessment and reminder of construction defects. Furthermore, by analyzing the fluctuations of grouting pressure and flow, combined with multi-spectral imaging results, it can more accurately identify hidden defects such as loose grouting and material stratification, thereby effectively improving the level of control over the quality of municipal road construction.

[0026] Furthermore, the similarity comparison module M200 is used to perform the following method: The three-dimensional point cloud data and the standard three-dimensional point cloud data are subjected to voxel gridding processing to generate multiple resolution voxel grids; a point cloud feature descriptor is determined for each resolution voxel grid in the multiple resolution voxel grids, wherein the point cloud feature descriptor includes at least one of a normal direction, a curvature, and a local point density; and a multi-scale similarity comparison is performed at different scales using the point cloud feature descriptors.

[0027] Specifically, voxel gridding is a method of discretizing continuous three-dimensional space into regular cubes (voxels), which is used to layer and simplify three-dimensional point cloud data; voxel grids of different resolutions refer to voxelization of three-dimensional point cloud data at different scales, with high-resolution grids providing detailed local features and low-resolution grids providing overall structural information; point cloud feature descriptors are parameters used to describe the local geometric characteristics of point clouds, including normal direction (indicating surface direction), curvature (indicating the degree of surface curvature) and local point density (indicating the distribution density of point clouds in local areas); multi-scale similarity comparison is to quantify the degree of similarity between the target point cloud and the standard point cloud by comparing their feature descriptors at different resolutions.

[0028] Execution steps: First, the three-dimensional point cloud data obtained by scanning the inspection and monitoring equipment is voxel-gridized with the standard three-dimensional point cloud data to generate multiple resolution voxel grids. Specifically, at a voxel side length of 1 cm, the fine geometric features of the target construction area can be captured, while at a side length of 10 cm, the focus is on the overall structural features; then, the point cloud feature descriptors under each resolution voxel grid are determined, such as the normal direction, curvature and local point density. By calculating the normal direction of the points within each voxel, the directional distribution of the surface of the construction area can be obtained, and the local point density can reflect the density of the point cloud.

[0029] Using these feature descriptors, multi-scale similarity comparison is performed at different scales. By comparing the similarity of the normal direction between the target and the standard point cloud at different scale resolutions, the deviation of the local surface direction and the difference in the overall structure are identified, effectively improving the recognition rate of minor defects.

[0030] Furthermore, the similarity comparison module M200 is further configured to perform the following method: The multi-scale similarity comparison indicators include any one or more of point pair distance, normal vector consistency, and feature point distribution density.

[0031] Specifically, multi-scale similarity comparison quantifies the degree of similarity between the three-dimensional point cloud data of the target construction area and the standard three-dimensional point cloud data by performing feature analysis on voxel grids of different resolutions; point-to-point distance refers to the Euclidean distance between the corresponding points in the target point cloud and the standard point cloud at the same or different scales, which is used to measure the deviation between the positions of the points; normal vector consistency determines whether the surface orientation is consistent by comparing the normal vector directions of the corresponding points in the target and standard point clouds; feature point distribution density refers to the distribution of the number of key feature points in a specific area, which is used to evaluate the richness of details in the point cloud data.

[0032] Execution steps: When performing multi-scale similarity comparison, first calculate the point-to-point distance between the target and the standard point cloud under different resolution voxel grids. Through the point-to-point distance between the target and the standard point cloud under the resolution voxel grid, it is found that the average distance exceeds the preset threshold, indicating that there is a significant geometric deviation in the target construction area; then, analyze the consistency of the normal vector by calculating the angle between the normal vectors of the corresponding points of the target and the standard point cloud. If the angle exceeds the set threshold, it indicates that there is a difference in the surface direction; finally, evaluate the distribution density of feature points by comparing the number of feature points in a specific area between the target and the standard point cloud. If the density of feature points of the target point cloud is significantly lower than that of the standard point cloud, it indicates that there is a certain probability that there are problems such as missing materials or loose structure in the area.

[0033] By combining indicators such as point-to-point distance, normal vector consistency, and feature point distribution density, hidden defects such as loose grouting and material stratification can be identified more accurately, effectively improving the quality control level of municipal road construction. By conducting multi-scale similarity comparison in municipal road construction quality inspection, construction quality can be comprehensively evaluated from multiple angles.

[0034] Furthermore, the association mapping module M300 is used to perform the following method: Capture an RGB image of the target construction area, and extract color features, edge features, and texture frequency features from the RGB image as first texture features; capture a near-infrared image of the target construction area, and extract moisture content features, material density features, and hidden defect features from the near-infrared image as second texture features.

[0035] Specifically, RGB images refer to images composed of three color channels: red, green, and blue, which can capture the surface color and texture information of the target construction area; color features are used to identify the characteristics of different materials or defects by analyzing the color distribution of each area in the image; edge features are used to identify structural boundaries by detecting areas with obvious changes in grayscale values ​​in the image; texture frequency features are used to identify surface roughness and other characteristics by analyzing the repetition frequency of texture patterns in the image.

[0036] Near-infrared imaging refers to imaging using light in the near-infrared band, which can penetrate the surface and detect internal information that cannot be detected by traditional visible light; moisture content characteristics identify the moisture content in the material through the characteristics of the interaction between near-infrared light and water molecules, material density characteristics identify the compactness of the material through the difference in reflection of near-infrared light in materials of different densities, and hidden defect characteristics identify hidden defects such as internal voids through the sensitivity of near-infrared light to internal structure.

[0037] Implementation steps: Use multispectral imaging equipment to capture RGB and near-infrared images of the target construction area. For the RGB images, use image processing algorithms to extract color, edge, and texture frequency features as the primary texture feature. Color features can be obtained by calculating the color histogram of each region in the image, edge features can be detected using Sobel or Canny operators, and texture frequency features can be analyzed using Gabor filters. These features can help identify visible defects such as surface cracks and material color differences.

[0038] For near-infrared images, moisture content, material density, and hidden defect features are extracted as secondary texture features. Moisture content can be calculated from the reflectance of a specific wavelength band, material density can be inferred from the image's grayscale distribution, and hidden defect features can be identified through unusual reflection patterns within the image. By fusing RGB and near-infrared image features, hidden defects such as internal voids and moisture penetration can be detected, enabling a more comprehensive assessment of construction quality and ensuring the accuracy and reliability of defect identification.

[0039] Furthermore, the defect reminder module M500 is used to perform the following method: A construction parameter state space is constructed using the grouting pump operating parameters and the grouting pump operating parameters; defect fuzzy probability is derived in the construction parameter state space to obtain the construction defect type and the defect fuzzy probability value; and based on the construction defect type and the defect fuzzy probability value, it is determined whether the construction defect reminder is triggered.

[0040] Specifically, the operating parameters of grouting pumps and grouting pumps refer to multiple indicators during equipment operation, such as grouting time, grouting pressure, grouting flow, slurry ratio, etc. Constructing the construction parameter state space is to integrate these parameters to form a multi-dimensional data space, which is used to comprehensively describe the state of the construction process. The derivation of defect fuzzy probability is to use fuzzy mathematics theory to calculate the possibility and type of construction defects based on the data in the construction parameter state space. The construction defect reminder is an early warning mechanism, which is triggered when the calculated defect fuzzy probability value exceeds the set threshold, reminding construction personnel to pay attention to and deal with potential defects.

[0041] Implementation steps: First, collect the operating parameters of the anchor maintenance point grouting and grouting construction process, such as grouting time, grouting pressure, grouting flow, etc. The construction parameter state space is constructed based on the grouting pump operating parameters and the grouting pump operating parameters, providing a data basis for defect analysis. Preferably, the grouting pressure parameter can reflect the density of the slurry filling. If the grouting pressure is abnormal, it indicates that the grouting is not dense. The grouting time parameter can reflect the adequacy of the grouting process. If the grouting time is too short, it indicates that the grouting is insufficient.

[0042] The defect fuzzy probability is deduced in the construction parameter state space to obtain the construction defect type and defect fuzzy probability value. Specifically, the fuzzy mathematical model is used to deduce the fuzzy probability of cavity defects in the target area based on parameters such as grouting pressure and grouting time. Based on the construction defect type and defect fuzzy probability value, it is determined whether a construction defect reminder is triggered, and construction personnel are warned in time so that corresponding measures can be taken, such as re-grouting or strengthening grouting, effectively improving the reliability and safety of construction quality.

[0043] Furthermore, the defect reminder module M500 is further configured to execute the following method: Determine the grouting pump operating parameters corresponding to the first grouting construction of the anchor maintenance point, including grouting time, grouting pressure, grouting flow, and slurry ratio; determine the grouting pump operating parameters corresponding to the second grouting construction of the anchor maintenance point, including grouting time, grouting pressure peak, pressure stabilization time, and grouting volume.

[0044] Specifically, anchor maintenance points are key locations for monitoring and evaluation in municipal road construction. The operating parameters of grouting pumps and grouting pumps correspond to the technical indicators in the primary grouting construction and the secondary grouting construction respectively; the operating parameters of the grouting pumps in the primary grouting construction include grouting time, grouting pressure, grouting flow, and slurry ratio, which respectively represent the duration of the grouting process, the applied pressure, the flow rate of the slurry, and the composition ratio of the slurry; the operating parameters of the grouting pumps in the secondary grouting construction include grouting time, grouting pressure peak, pressure stabilization time, and grouting volume, which respectively represent the duration of the grouting process, the maximum pressure value reached, the time period for maintaining stable pressure, and the total amount of slurry pressed in, which together reflect the quality and status of grouting and grouting construction.

[0045] Execution steps: First, determine the grouting pump operating parameters corresponding to the primary grouting construction of the anchor maintenance point, including grouting time, grouting pressure, grouting flow, and slurry ratio. At the same time, determine the grouting pump operating parameters corresponding to the secondary grouting construction, including grouting time, grouting pressure peak, pressure stabilization time, and grouting volume. Generally speaking, insufficient grouting time has a certain probability of causing the slurry to fail to be fully filled, insufficient grouting pressure leads to loose grouting, too small grouting flow leads to low grouting efficiency, and improper slurry ratio may cause the slurry performance to not meet the requirements. Insufficient grouting time has a certain probability of causing insufficient grouting, too low a grouting pressure peak has a certain probability of causing poor grouting effect, insufficient pressure stabilization time leads to poor stability after grouting, and insufficient grouting volume leads to insufficient density after grouting.

[0046] These parameters are used to deduce the fuzzy probability of defects through the construction parameter state space, which affects the triggering of construction defect reminders. Through the analysis of the operating parameters of the grouting pump and the grouting pump, it was found that the grouting pressure of a grouting operation was lower than the normal range. Combined with other parameters, the fuzzy probability of the cavity defect is derived, which triggers the construction defect reminder and prompts the construction personnel to deal with it, effectively improving the reliability and safety of the construction quality.

[0047] Furthermore, the defect reminder module M500 is further configured to execute the following method: If the construction defect reminder is triggered, the defect range is preliminarily defined through multi-scale similarity comparison information; the boundary area of ​​the preliminarily defined defect range is cropped based on the first texture feature corresponding to the visible light band; and the boundary area of ​​the preliminarily defined defect range is cropped based on the second texture feature corresponding to the near-infrared band.

[0048] Specifically, multi-scale similarity comparison information is the difference information obtained by comparing the three-dimensional point cloud data of the target construction area with the standard three-dimensional point cloud data at different resolutions, which is used to preliminarily determine the possible scope of defects; the first texture features corresponding to the visible light band include color features, edge features and texture frequency features. These features can clearly outline the details of the surface of the construction area and provide an explicit basis for boundary clipping; the second texture features corresponding to the near-infrared band cover moisture content features, material density features and hidden defect features, which can reveal potential problems inside the construction area and assist in determining the defect boundary from an implicit perspective.

[0049] Execution steps: First, use multi-scale similarity comparison information to quickly locate areas where defects may exist and obtain a preliminary definition of the defect range. Through comparison, it is found that there is a significant difference between a certain local area of ​​the target construction area and the standard model, and it is preliminarily judged that there may be defects in this area; next, use the first texture feature corresponding to the visible light band to crop the boundary area of ​​the preliminarily defined defect range.

[0050] By analyzing the color, edge, and texture frequency features in the RGB image, the defect boundary can be accurately identified. Specifically, color features can help distinguish different materials or identify color differences, while edge features can clearly outline the defect. Secondary texture features corresponding to the near-infrared band are then used to further optimize the cropping of defect boundaries.

[0051] By analyzing the moisture content characteristics, material density characteristics and hidden defect characteristics in near-infrared images, potential problems inside the defects are revealed and the accuracy of boundary clipping is ensured. Specifically, the moisture content characteristics can help identify internal voids caused by moisture penetration, and the material density characteristics can reflect the compactness of the material. Through multi-dimensional feature fusion analysis, the location and scope of the defect can be determined more accurately.

[0052] Furthermore, the defect reminder module M500 is further configured to execute the following method: Superpixel segmentation is performed on the RGB image within the preliminarily defined defect range, and the image is divided into multiple superpixel blocks with similar color features; the color histogram features, local binary pattern features and grayscale co-occurrence matrix features of multiple superpixel blocks are extracted to construct the feature vector of the superpixel block; the boundary area of ​​the preliminarily defined defect range is cropped based on the feature vector of the superpixel block and the anisotropic characteristics of the edge of the defect area.

[0053] Specifically, superpixel segmentation refers to dividing an image into multiple superpixel blocks with similar color features. These superpixel blocks can better reflect the local features and structural information of the image than traditional pixels; the color histogram feature describes the distribution of colors in the image, the local binary pattern feature is used to capture the local texture pattern of the image, and the grayscale co-occurrence matrix feature is used to quantify the spatial relationship and texture characteristics of the grayscale values ​​in the image; the anisotropy feature reflects the directionality and non-uniformity of the edge of the defect area, and is used to evaluate the complexity and irregularity of the edge.

[0054] Execution steps: within the initially defined defect range, first perform superpixel segmentation on the RGB image to divide the image into multiple superpixel blocks with similar color features; then, extract the color histogram features, local binary pattern features, and grayscale co-occurrence matrix features of each superpixel block to construct the feature vector of each superpixel block.

[0055] Color histogram features can help identify the distribution of different color areas, local binary pattern features can capture edge and texture details, and grayscale co-occurrence matrix features can quantify the roughness and directionality of the texture; finally, the boundary area of ​​the defect range is preliminarily defined through the feature vector of the superpixel block and the anisotropic characteristics of the edge of the defect area. Specifically, the cosine similarity between the feature vector of the superpixel block and the feature vector of the defect area, as well as the anisotropic characteristics of the edge of the defect area, are used to accurately identify and crop the defect boundary area, effectively improving the accuracy of defect boundary identification, ensuring the accuracy of defect positioning, and providing reliable data support for subsequent defect repair.

[0056] Furthermore, the defect reminder module M500 is further configured to execute the following method: Based on the cosine similarity between the feature vector of the superpixel block and the feature vector of the defect area, a similarity threshold τ is defined; multiple superpixel blocks are traversed, and superpixel blocks with cosine similarity less than the similarity threshold τ are removed from the preliminary defined defect range; the edge of the cropped defect area is extracted; the corner points and mutation points on the edge of the cropped defect area are identified, the normal vector field of the boundary pixels is generated, the angle distribution between the normal vector and the local texture direction is determined, and the anisotropic characteristics of the edge of the defect area are evaluated.

[0057] Specifically, the superpixel block feature vector is constructed by extracting the color histogram features, local binary pattern features and gray-level co-occurrence matrix features of the superpixel block, which is used to describe the visual characteristics of the superpixel block. The defect area feature vector is obtained based on the feature statistics of the known defect area and is used for comparison with the superpixel block feature vector.

[0058] Cosine similarity is an indicator that measures the degree of similarity between two vectors. Its value range is between 0 and 1. The closer the value is to 1, the higher the similarity. The similarity threshold τ is a pre-set judgment standard used to distinguish whether a superpixel block belongs to a defect area. The anisotropic feature reflects the directionality and non-uniformity of the edge of the defect area and is used to evaluate the complexity and irregularity of the edge. The normal vector field is a vector field that describes the direction of the boundary pixel, and the local texture direction is the main direction of the texture near the boundary pixel.

[0059] Execution steps: First, define a similarity threshold τ based on the cosine similarity between the feature vector of the superpixel block and the feature vector of the defect area. Specifically, through training with a large amount of sample data, it is determined that when the cosine similarity is less than 0.6, the superpixel block is likely to belong to a non-defective area; then, traverse multiple superpixel blocks within the initially defined defect range and calculate the cosine similarity between each superpixel block and the feature vector of the defect area; remove superpixel blocks with a cosine similarity less than the similarity threshold τ from the initially defined defect range, thereby more accurately narrowing the defect range and making the defect area more prominent.

[0060] Next, the edges of the cropped defect area are extracted, and the corner points and mutation points on the edges are identified. These key points usually indicate the starting or ending position of the defect and are crucial for the accurate positioning of the defect. Then, the normal vector field of the boundary pixels is generated, and the anisotropic characteristics of the edge of the defect area are evaluated by calculating the angle distribution between the normal vector and the local texture direction. Specifically, if the angle distribution between the normal vector and the local texture direction presents a Gaussian distribution, it indicates that the defect edge is relatively regular; if the angle distribution is relatively discrete, it indicates that the defect edge is relatively complex. The analysis method based on feature vectors and anisotropic characteristics can effectively improve the accuracy of defect boundary recognition and ensure the accuracy of defect positioning.

[0061] In summary, the beneficial effects of the embodiments of the present application are: Due to the use of a scanning module, the inspection and monitoring equipment is set at the first position coordinate of the associated mapping of the anchor maintenance point; the inspection and monitoring equipment is used to scan the target construction area of ​​the municipal road; a similarity comparison module is used to perform multi-scale similarity comparison between the three-dimensional point cloud data obtained by the inspection and monitoring equipment scanning and the standard three-dimensional point cloud data stored in the historical qualified samples; an associated mapping module is used to set the multi-spectral imaging equipment at the second position coordinate of the associated mapping of the anchor maintenance point; the multi-spectral imaging equipment is used to synchronously obtain the first texture features corresponding to the visible light band and the second texture features corresponding to the near-infrared band of the target construction area; a parameter search module is used to search for the grouting pump operating parameters of the anchor maintenance point during the first grouting construction process; the grouting pump operating parameters of the anchor maintenance point during the second grouting construction process are searched; a defect reminder module is used to remind of construction defects based on the multi-scale similarity comparison information, the first texture features corresponding to the visible light band, the second texture features corresponding to the near-infrared band, combined with the grouting pump operating parameters and the grouting pump operating parameters. This application provides a municipal road construction defect inspection system and method based on visual inspection, which realizes the multi-scale similarity comparison analysis of the differences between the target construction area and the standard model, more comprehensively detects structural defects, and integrates and analyzes visual inspection data with construction process parameters, comprehensively evaluates construction quality from multiple angles, and determines the specific location of defects in space to provide defect positioning reminders. Technical effect.

[0062] Example 2 Based on the same inventive concept as the municipal road construction defect inspection system based on visual detection in the aforementioned embodiment, Figure 2 As shown, an embodiment of the present application provides a municipal road construction defect inspection method based on visual detection, wherein the method includes: S1: The inspection and monitoring equipment is set at the first position coordinate associated with the anchor maintenance point; and the inspection and monitoring equipment is used to scan the target construction area of ​​the municipal road.

[0063] S2: Perform a multi-scale similarity comparison between the three-dimensional point cloud data obtained by scanning the inspection and monitoring equipment and the standard three-dimensional point cloud data stored in historical qualified samples.

[0064] S3: A multispectral imaging device is set at the second position coordinate associated with the anchor maintenance point; the multispectral imaging device is used to synchronously obtain a first texture feature corresponding to the visible light band and a second texture feature corresponding to the near-infrared band of the target construction area; S4: searching for the operating parameters of the grouting pump during the primary grouting construction process of the anchor maintenance point; searching for the operating parameters of the grouting pump during the secondary grouting construction process of the anchor maintenance point.

[0065] S5: Based on the multi-scale similarity comparison information, the first texture feature corresponding to the visible light band, the second texture feature corresponding to the near-infrared band, and the grouting pump operating parameters and the grouting pump operating parameters, construction defect reminders are issued.

[0066] Furthermore, a multi-scale similarity comparison is performed on the three-dimensional point cloud data scanned by the inspection and monitoring equipment and the standard three-dimensional point cloud data stored in the historical qualified samples. The method of the present application includes: The three-dimensional point cloud data and the standard three-dimensional point cloud data are subjected to voxel gridding processing to generate multiple resolution voxel grids; a point cloud feature descriptor is determined for each resolution voxel grid in the multiple resolution voxel grids, wherein the point cloud feature descriptor includes at least one of a normal direction, a curvature, and a local point density; and a multi-scale similarity comparison is performed at different scales using the point cloud feature descriptors.

[0067] Furthermore, the method of the present application includes: performing multi-scale similarity comparison at different scales using the point cloud feature descriptor; The multi-scale similarity comparison indicators include any one or more of point pair distance, normal vector consistency, and feature point distribution density.

[0068] Furthermore, the method of the present application includes: synchronously acquiring a first texture feature corresponding to the visible light band and a second texture feature corresponding to the near-infrared band of the target construction area; Capture an RGB image of the target construction area, and extract color features, edge features, and texture frequency features from the RGB image as first texture features; capture a near-infrared image of the target construction area, and extract moisture content features, material density features, and hidden defect features from the near-infrared image as second texture features.

[0069] Furthermore, the construction defect reminder is performed in combination with the grouting pump operating parameters and the grouting pump operating parameters. The application method includes: A construction parameter state space is constructed using the grouting pump operating parameters and the grouting pump operating parameters; defect fuzzy probability is derived in the construction parameter state space to obtain the construction defect type and the defect fuzzy probability value; and based on the construction defect type and the defect fuzzy probability value, it is determined whether the construction defect reminder is triggered.

[0070] Furthermore, the present application method includes: Determine the grouting pump operating parameters corresponding to the first grouting construction of the anchor maintenance point, including grouting time, grouting pressure, grouting flow, and slurry ratio; determine the grouting pump operating parameters corresponding to the second grouting construction of the anchor maintenance point, including grouting time, grouting pressure peak, pressure stabilization time, and grouting volume.

[0071] Furthermore, after determining whether the construction defect reminder is triggered, the present application method further includes: If the construction defect reminder is triggered, the defect range is preliminarily defined through multi-scale similarity comparison information; the boundary area of ​​the preliminarily defined defect range is cropped based on the first texture feature corresponding to the visible light band; and the boundary area of ​​the preliminarily defined defect range is cropped based on the second texture feature corresponding to the near-infrared band.

[0072] Furthermore, based on the first texture feature corresponding to the visible light band, the boundary area of ​​the initially defined defect range is cropped. The method of the present application includes: Superpixel segmentation is performed on the RGB image within the preliminarily defined defect range, and the image is divided into multiple superpixel blocks with similar color features; the color histogram features, local binary pattern features and grayscale co-occurrence matrix features of multiple superpixel blocks are extracted to construct the feature vector of the superpixel block; the boundary area of ​​the preliminarily defined defect range is cropped based on the feature vector of the superpixel block and the anisotropic characteristics of the edge of the defect area.

[0073] Furthermore, the present application method includes: Based on the cosine similarity between the feature vector of the superpixel block and the feature vector of the defect area, a similarity threshold τ is defined; multiple superpixel blocks are traversed, and superpixel blocks with cosine similarity less than the similarity threshold τ are removed from the preliminary defined defect range; the edge of the cropped defect area is extracted; the corner points and mutation points on the edge of the cropped defect area are identified, the normal vector field of the boundary pixels is generated, the angle distribution between the normal vector and the local texture direction is determined, and the anisotropic characteristics of the edge of the defect area are evaluated.

[0074] In summary, any step can be stored as a computer instruction or program in an unlimited computer memory and can be called and recognized by an unlimited computer processor, without any unnecessary restrictions.

[0075] Furthermore, the above technical solution only reflects the preferred technical solution of the technical solution of the embodiment of the present application. Some changes that may be made to certain parts thereof by technical personnel in this technical field all reflect the novel principles of the embodiment of the present application. Obviously, technical personnel in this field can make various changes and modifications to the present application without departing from the scope of the present application.

Claims

1. A municipal road construction defect inspection system based on visual inspection, characterized in that: The system comprises: A scanning module is used to set the patrol monitoring device at the first position coordinate associated with the anchor maintenance point; and use the patrol monitoring device to scan the target construction area of ​​the municipal road; A similarity comparison module is used to perform multi-scale similarity comparison between the three-dimensional point cloud data scanned by the inspection and monitoring equipment and the standard three-dimensional point cloud data stored in historical qualified samples; An association mapping module is used to set a multispectral imaging device at a second position coordinate associated with the anchor maintenance point; using the multispectral imaging device, synchronously obtain a first texture feature corresponding to a visible light band and a second texture feature corresponding to a near-infrared band of the target construction area; A parameter search module is used to search for the operating parameters of the grouting pump of the anchor maintenance point during the first grouting construction process; and to search for the operating parameters of the grouting pump of the anchor maintenance point during the second grouting construction process; The defect reminder module is used to provide construction defect reminders based on multi-scale similarity comparison information, the first texture features corresponding to the visible light band, the second texture features corresponding to the near-infrared band, and the grouting pump operating parameters and the grouting pump operating parameters.

2. A municipal road construction defect inspection system based on visual inspection as claimed in claim 1, characterized in that: Performing a multi-scale similarity comparison on the three-dimensional point cloud data scanned by the inspection and monitoring equipment and the standard three-dimensional point cloud data stored in historical qualified samples, including: Performing voxel gridding processing on the three-dimensional point cloud data and standard three-dimensional point cloud data to generate voxel grids of multiple resolutions; Determine a point cloud feature descriptor for each resolution voxel grid in the plurality of resolution voxel grids, wherein the point cloud feature descriptor comprises at least one of a normal direction, a curvature, and a local point density; Multi-scale similarity comparison is performed at different scales using the point cloud feature descriptor.

3. A municipal road construction defect inspection system based on visual inspection as claimed in claim 2, characterized in that: Performing multi-scale similarity comparison at different scales using the point cloud feature descriptor, including: The multi-scale similarity comparison indicators include any one or more of point pair distance, normal vector consistency, and feature point distribution density.

4. A municipal road construction defect inspection system based on visual inspection as claimed in claim 1, characterized in that: Synchronously obtain the first texture feature corresponding to the visible light band and the second texture feature corresponding to the near-infrared band of the target construction area, including: Capturing an RGB image of the target construction area, and extracting color features, edge features, and texture frequency features from the RGB image as first texture features; A near-infrared image of the target construction area is captured, and moisture content features, material density features, and hidden defect features in the near-infrared image are extracted as second texture features.

5. A municipal road construction defect inspection system based on visual inspection as claimed in claim 1, characterized in that: Combined with the grouting pump operating parameters and grouting pump operating parameters, construction defect reminders are provided, including: The construction parameter state space is constructed through the grouting pump operating parameters and the grouting pump operating parameters; Derivation of defect fuzzy probability in the construction parameter state space is performed to obtain the construction defect type and the defect fuzzy probability value; Whether to trigger the construction defect reminder is determined according to the construction defect type and the defect fuzzy probability value.

6. A municipal road construction defect inspection system based on visual inspection as claimed in claim 5, characterized in that: include: Determine the grouting pump operating parameters corresponding to the one-time grouting construction of the anchor curing point, including grouting time, grouting pressure, grouting flow rate, and slurry ratio; Determine the grouting pump operating parameters corresponding to the secondary grouting construction of the anchor maintenance point, including grouting time, grouting pressure peak, pressure stabilization time, and grouting volume.

7. A municipal road construction defect inspection system based on visual inspection as claimed in claim 6, characterized in that: After determining whether the construction defect reminder is triggered, the method further includes: If the construction defect reminder is triggered, the defect scope is preliminarily defined through multi-scale similarity comparison information; The boundary region of the initially defined defect range is cropped according to the first texture feature corresponding to the visible light band; and the boundary region of the initially defined defect range is cropped according to the second texture feature corresponding to the near-infrared band.

8. A municipal road construction defect inspection system based on visual inspection as claimed in claim 7, characterized in that: The boundary region of the initially defined defect range is cropped according to the first texture feature corresponding to the visible light band, including: Performing superpixel segmentation on the RGB image within the initially defined defect range to divide the image into multiple superpixel blocks with similar color features; Extract the color histogram features, local binary pattern features, and gray-level co-occurrence matrix features of multiple superpixel blocks to construct the feature vector of the superpixel block; The boundary area of ​​the initially defined defect range is cropped using the feature vector of the superpixel block and the anisotropic characteristics of the edge of the defect area.

9. A municipal road construction defect inspection system based on visual inspection as claimed in claim 8, characterized in that: include: According to the cosine similarity between the super pixel block feature vector and the defect area feature vector, a similarity threshold τ is defined; Traversing multiple superpixel blocks, removing superpixel blocks whose cosine similarity is less than a similarity threshold τ from the initially defined defect range; extracting the edge of the cropped defect area; Identify the corner points and mutation points on the edge of the cropped defect area, generate the normal vector field of the boundary pixels, determine the angle distribution between the normal vector and the local texture direction, and evaluate the anisotropic characteristics of the edge of the defect area.

10. A method for inspecting defects in municipal road construction based on visual inspection, characterized in that: A method for implementing a municipal road construction defect inspection system based on visual inspection according to any one of claims 1 to 9, comprising: The inspection and monitoring device is set at the first position coordinate associated with the anchor maintenance point; the inspection and monitoring device is used to scan the target construction area of ​​the municipal road; Performing a multi-scale similarity comparison between the three-dimensional point cloud data obtained by scanning the inspection and monitoring equipment and the standard three-dimensional point cloud data stored in historical qualified samples; The multispectral imaging device is set at the second position coordinate associated with the anchor maintenance point; the multispectral imaging device is used to synchronously obtain the first texture feature corresponding to the visible light band and the second texture feature corresponding to the near infrared band of the target construction area; Searching for the operating parameters of the grouting pump at the anchor maintenance point during the primary grouting construction process; searching for the operating parameters of the grouting pump at the anchor maintenance point during the secondary grouting construction process; Based on multi-scale similarity comparison information, the first texture features corresponding to the visible light band, the second texture features corresponding to the near-infrared band, and combined with the grouting pump operating parameters and the grouting pump operating parameters, construction defect reminders are given.

Images