CN-121169876-B - Expressway construction inspection optimization method based on artificial intelligence

CN121169876BCN 121169876 BCN121169876 BCN 121169876BCN-121169876-B

Abstract

The invention relates to the technical field of construction inspection and discloses an expressway construction inspection optimization method based on artificial intelligence, which comprises the steps of adopting a boundary recognition model to recognize foundation pit boundary pixels in a construction ground image, mapping the foundation pit boundary pixels to an expressway construction site, and taking an area surrounded by mapping positions of the foundation pit boundary pixels as a foundation pit; and generating a security risk value of the discrete grid based on the foundation pit information of the foundation pit in the discrete grid, and performing secondary inspection on the discrete grid with the security risk value higher than the allowable risk. According to the invention, the depth, width and area of the foundation pit are obtained through a stereo matching algorithm, and a security risk value is generated by combining the discrete grids, so that the accurate secondary inspection of the high-risk discrete grids and the accurate security management of expressway construction are realized.

Inventors

LI QINGHAO
LIU YUGUO
SUN HAIDONG

Assignees

聊城市交通发展有限公司

Dates

Publication Date: 20260512
Application Date: 20250918

Claims (7)

1. The highway construction inspection optimization method based on artificial intelligence is characterized by comprising the following steps of: S1, acquiring a construction ground image of a highway construction site by using a depth camera at a fixed position, identifying foundation pit boundary pixels in the construction ground image by using a boundary identification model, mapping the foundation pit boundary pixels to the highway construction site, and taking an area surrounded by the mapping position of the foundation pit boundary pixels as a foundation pit; s2, controlling a laser radar scanner deployed on the mobile inspection vehicle to acquire auxiliary point cloud data of the foundation pit, and controlling a binocular camera deployed on the mobile inspection vehicle to acquire a left view and a right view of the foundation pit; S3, calculating parallax information of the foundation pit based on the left view and the right view, and estimating the depth and the width of the foundation pit by adopting a three-dimensional matching algorithm in combination with auxiliary point cloud data to obtain the depth and the width of the foundation pit; S4, dividing a highway construction site into discrete grid areas, generating a security risk value of the discrete grid based on foundation pit information of a foundation pit in the discrete grid, and performing secondary inspection on the discrete grid with the security risk value higher than an allowable risk, wherein the foundation pit information comprises the area, depth and width of the foundation pit; the security risk value calculation formula of the discrete grid is as follows: ; wherein R represents a security risk value of the discrete grid, len represents a side length of the discrete grid, H represents the number of foundation pits in the discrete grid, Representing the area of the h pit in the discrete grid, Representing the depth of the h foundation pit within the discrete grid, Representing the width of the h pit in the discrete grid, Representing the maximum depth of all pit in the highway construction site, Respectively in turn is area Depth of Width of the container Is used for the normalization of the values of (c), Representing the volume duty factor of the h foundation pit; all of which represent the coefficient of effect, Representing the nonlinear amplification factor.
2. The method for optimizing highway construction inspection based on artificial intelligence according to claim 1, wherein the step of identifying the boundary pixels of the foundation pit in the construction ground image by using the boundary identification model comprises the steps of: Deploying a plurality of depth cameras at fixed positions on the expressway construction site, acquiring construction ground images of the expressway construction site by using the deployed cameras, and transmitting the acquired construction ground images to a boundary recognition model; the boundary recognition model comprises an input layer, a contrast enhancement layer and a boundary recognition layer, wherein the input layer is used for receiving a construction ground image and carrying out graying treatment to obtain a construction ground gray image, the contrast enhancement layer is used for calculating the contrast of the construction ground gray image, the contrast enhancement treatment is carried out on the construction ground gray image with low contrast by adopting a histogram equalization method, gamma correction treatment is carried out on the construction ground gray image with overexposed or underexposed to obtain a construction ground contrast enhancement image, the boundary recognition layer is used for carrying out foundation pit boundary pixel recognition on the construction ground contrast enhancement image to obtain foundation pit boundary pixels in the construction ground contrast enhancement image, and pixel coordinates of the foundation pit boundary pixels are extracted; the boundary recognition layer adopts an improved YOLOv model structure, and the improvement mode is that a multi-scale characteristic fusion structure is introduced into a main network; And carrying out boundary recognition on the construction ground image by using the boundary recognition model to obtain pixel coordinates of pixels recognized as the foundation pit boundary, and mapping the foundation pit boundary pixels to the expressway construction site by using the pixel coordinates of the foundation pit boundary pixels.
3. The method for optimizing the inspection tour of highway construction based on artificial intelligence according to claim 2, wherein the mapping the boundary pixels of the foundation pit to the construction site of the highway by using the pixel coordinates of the boundary pixels of the foundation pit and taking the area surrounded by the mapping positions of the boundary pixels of the foundation pit as the foundation pit comprises: Extracting a depth value at a pixel coordinate of a foundation pit boundary pixel, and mapping the pixel coordinate to a camera coordinate system based on an internal reference matrix of a depth camera by combining the depth value to obtain a three-dimensional camera coordinate of the pixel coordinate of the foundation pit boundary pixel in the camera coordinate system; Converting the three-dimensional camera coordinates into a unified expressway coordinate system by using an external parameter rotation matrix and an external parameter translation vector of the depth camera to obtain three-dimensional expressway coordinates of the three-dimensional camera coordinates in the expressway coordinate system; And connecting the three-dimensional highway coordinates in a highway coordinate system to form a closed polygon, wherein the closed polygon is an area surrounded by the pixel mapping positions of the foundation pit boundary.
4. The method for optimizing highway construction inspection based on artificial intelligence according to claim 1, wherein controlling the laser radar scanner deployed on the mobile inspection vehicle to collect auxiliary point cloud data of the foundation pit and controlling the binocular camera deployed on the mobile inspection vehicle to collect left and right views of the foundation pit comprises: in the expressway construction site, the movable inspection vehicle is deployed at a proper position around the foundation pit, so that the movable inspection vehicle can cover each foundation pit on an inspection path; The mobile inspection vehicle is provided with a laser radar scanner and a binocular camera, wherein the binocular camera comprises a left view camera and a right view camera which are respectively used for collecting a left view and a right view; Controlling the mobile inspection vehicle to move along the periphery of the foundation pit, enabling the laser radar scanner to perform 360-degree rotation scanning on the foundation pit at a preset scanning frequency, acquiring three-dimensional point clouds of the foundation pit and the periphery of the foundation pit, forming three-dimensional point clouds to serve as auxiliary point cloud data of the foundation pit, and enabling the three-dimensional point clouds to be in a three-dimensional coordinate form; and controlling the mobile inspection vehicle to synchronously start shooting of the left view camera and the right view camera, and collecting a left view and a right view of the panorama of the foundation pit.
5. The method for optimizing highway construction inspection based on artificial intelligence according to claim 4, wherein the parallax information of the foundation pit is calculated based on the left view and the right view, comprises: carrying out graying treatment on the left view and the right view to obtain a graying left view and a graying right view; Setting the parallax search range as ; Calculating parallax values of any pixel coordinates in the grayscale left view under different parallaxes, wherein the parallax values are used as parallax information of a foundation pit: ; Wherein, the Represents the parallax value of the pixel coordinate u in the grayed-out left view at the parallax q, A set of pixel coordinates representing a greyscale left view, Representing the center of pixel coordinate u in a greyscale left view Pixel region, v denotes pixel region Is used to determine the pixel coordinates of the pixel, Representing the gray value at pixel coordinate v in the grayed-out left view, Representing pixel coordinates after the pixel coordinates v are translated q pixels in the horizontal direction, Representing pixel coordinates in a grayed right view Gray values at; And acquiring auxiliary point cloud data of the foundation pit.
6. The method for optimizing highway construction inspection based on artificial intelligence according to claim 5, wherein the estimating the depth and the width of the foundation pit by adopting a stereo matching algorithm in combination with auxiliary point cloud data to obtain the depth and the width of the foundation pit comprises: Calculating to obtain a point cloud consistent matching cost between any pixel coordinate in the grayscale left view and auxiliary point cloud data, wherein the point cloud consistent matching cost is the minimum coordinate difference between the pixel coordinate and three-dimensional point cloud projection results of all three-dimensional point clouds in the auxiliary point cloud data, and the three-dimensional point cloud projection results are that the three-dimensional point cloud projects to a pixel coordinate system where the pixel coordinate in the grayscale left view is located; Adding the point cloud consistent matching cost between the pixel coordinates in the grayscale left view and the auxiliary point cloud data and the parallax values of the pixel coordinates under different parallaxes to serve as the initial matching cost of the pixel coordinates in the grayscale left view under different parallaxes; Carrying out cost accumulation in different directions on initial matching costs of pixel coordinates in the grayscale left view under different parallaxes, and summing cost accumulation results in multiple directions to obtain path accumulation costs of the pixel coordinates under different parallaxes; Selecting parallax for minimizing path accumulation cost of pixel coordinates in the grayscale left view, generating parallax depth of the pixel coordinates in the grayscale left view based on the selected parallax, a camera baseline and a focal length, taking the parallax depth as a depth value of the pixel coordinates in the grayscale left view, mapping the pixel coordinates in the grayscale left view to an expressway coordinate system by adopting a mode of mapping the pixel to an expressway construction site in the step S1 to obtain mapped three-dimensional highway coordinates, and taking the coordinate value of the mapped three-dimensional highway coordinates in the Z axis as the depth of the pixel coordinates in the position of a foundation pit in the grayscale left view; selecting mapped three-dimensional highway coordinates of all pixel coordinates in the grayscale left view, projecting the mapped three-dimensional highway coordinates into a plane coordinate system in a highway coordinate system to obtain projection coordinates of each mapped three-dimensional working coordinate, calculating Euclidean distance between any two projection coordinates, and selecting the maximum Euclidean distance as the width of a foundation pit; And selecting the maximum value of the depth of all pixel coordinates in the gray left view at the position of the foundation pit as the depth of the foundation pit.
7. The method for optimizing highway construction inspection according to claim 1, wherein step S4 comprises: projecting the highway construction site into a planar coordinate system within the highway coordinate system; dividing the projected highway construction site into a plurality of square discrete grids; and generating a security risk value of the discrete grid based on the foundation pit information of the foundation pit in the discrete grid.

Description

Expressway construction inspection optimization method based on artificial intelligence Technical Field The invention relates to the field of construction inspection, in particular to an image recognition technology, and specifically relates to an artificial intelligence-based expressway construction inspection optimization method. Background The construction and maintenance of the expressway are important basic links for guaranteeing long-term safe operation of roads and improving passing efficiency. Along with the continuous increase of traffic flow and the continuous increase of road use frequency, construction operation is gradually changed from traditional manual inspection to an intelligent and informationized direction. However, during construction and maintenance, foundation pits are a common construction structure that presents a non-negligible safety risk. On the one hand, the foundation pit is usually formed by excavation operation, the boundary and the shape of the foundation pit are often complex and changeable, if the foundation pit is not monitored in real time and accurately marked, the construction machinery is easy to be mistakenly put into or operators fall off, and serious personnel injury and equipment damage are caused. On the other hand, if the foundation pit is not recognized in time or the parameter evaluation is inaccurate, vehicles running near the construction area can be influenced by the problems of collapse of the foundation pit, insufficient support of the road surface and the like, so that traffic accidents are caused, and hidden danger is brought to the road safety and the running efficiency. The foundation pit not only has the difference in depth and width, but also the area and the space distribution position of the foundation pit can be dynamically adjusted along with the change of the construction progress. Therefore, the traditional method relying on manual inspection and experience judgment is difficult to ensure comprehensive grasp and accurate update of foundation pit information, and has the problems of insufficient monitoring delay and recognition accuracy. In an actual construction scene, environmental factors such as night work, complex terrain, rainwater accumulation and the like can also make boundary identification and risk judgment of a foundation pit more difficult. Therefore, how to realize rapid identification and high-precision measurement of the foundation pit area becomes a key technical bottleneck for restricting construction safety management. The existing research is focused on a highway inspection method based on artificial intelligence and big data, and patent CN117726324B provides an inspection mode combining a neural network and a prediction model for highway traffic construction inspection method and system based on data identification. According to the method, the maintenance value, the predicted vehicle transportation data, the predicted temperature data and the predicted precipitation data of a future time period are obtained, and the data are imported into a constructed neural network model to obtain the change trend of the future maintenance value. Further, the time for the maintenance value to reach the maintenance threshold is estimated, so that the prediction and the alarm of the road maintenance time are realized, the time difference between the road damage and the maintenance is reduced, and the prospective and the timeliness of the road maintenance are improved. Although the above patent advances in the aspects of road damage prediction and maintenance time early warning, the technical difficulty still exists in that the patent focuses on the prediction of time dimension, the quantitative measurement of spatial geometric features such as boundaries, depth, width and the like of a foundation pit cannot be performed, the warning is performed by predicting the maintenance time, the actual landform information and structural risks of a construction site cannot be combined, and the most dangerous foundation pit in construction is not subjected to targeted intelligent analysis. Aiming at the problem, the invention provides an artificial intelligence-based expressway construction inspection optimization method, which is used for carrying out automatic calculation measurement on a foundation pit on a construction site by introducing an artificial intelligence technology, realizing intelligent inspection, improving the inspection efficiency and precision, reducing the manual inspection cost and improving the overall management level of the construction process. Disclosure of Invention The invention provides an artificial intelligence-based highway construction inspection optimization method, wherein the traditional construction inspection is dependent on manual visual inspection or single camera equipment, is easy to be interfered by illumination change, dust shielding, weak texture and the like, so that the boundary detection of a foundation