3D Modeling And Repair Parameter Calculation Of Pavement Crack Based On Machine Vision

Nowadays,with the rapid development of highway construction,road detection and maintenance management has increasingly became an important task of highway construction in China.As an early manifestation of most pavement diseases,cracks are one of the most important parameters for evaluating pavement quality,which directly affects the service life of roads and driving safety.Early detection and repair can prevent further development of cracks and have serious consequences.At present,the detection and repair of pavement cracks is less intelligent and there is a large room for development.This paper proposes a road vision crack detection system based on machine vision.The depth image acquired by the Kinect camera is used to realize the three-dimensional reconstruction and parameter calculation of the crack.Firstly,the paper introduces the depth ranging principle of Kinect depth camera,and expounds the direct transformation relationship between image coordinate system,camera coordinate system and world coordinate system.The calibration of camera is realized by Matlab calibration toolbox and Zhang Zhengyou camera calibration method.Secondly,Analysis of the error in the image of the crack depth of the collected pavement,and the black void generated during the process of collecting the crack depth image of the pavement are eliminated.The noise in the depth image is processed by Gaussian filtering,median filtering,mean filtering and bilateral filtering,respectively.According to the characteristics of each filtering algorithm and the effect diagram after processing,the bilateral filtering is selected to carry out the edge preservation and denoising of the crack,and the method is verified and analyzed.The verification results show that the filtering algorithm has a good processing effect;then,before the 3D reconstruction,it needs to remove the uneven surface of the undamaged road surface,after the watershed segmentation algorithm is used for region segmentation,the image is masked to eliminate the influence of the non-destructive pavement,and then according to the relationship between the depth image and the coordinate system,the crack points are obtained.The three-dimensional coordinates of the world coordinate system are used to establish the three-dimensional visualization model of the crack.Finally,the length,maximum width,minimum circumscribed rectangle,maximum depth and average depth of the crack are calculated,and the actual measured and calculated parameters are calculated.Conduct a comparative verification analysis.The test results show that the proposed method based on machine vision for pavement crack detection has certain feasibility,and the establishment and detection of crack threedimensional visualization model is realized.Compared with the traditional pavement detection system,it can not only obtain the three-dimensional parameterized information of pavement cracks quickly,but also provide scientific and reliable technical support for crack repair.