Research On Aphalt Pavement Pothole Detection Technology Based On 3D Reconstruction Technology

Asphalt pavement diseases can decrease traffic safety and have a great impact on road structure.At present,the road condition assessment method is mainly manual and time-consuming.Existing methods for identifying and evaluating potholes rely on 3d surface reconstruction with high equipment and computational cost.As the total volume of highway in China is gradually increasing and the volume of heavy load traffic is increasing rapidly,the work of highway maintenance and management becomes more complicated and tedious.Comprehensive,reliable and fast methods of pothole indicators can prevent road safety accidents,improve highway maintenance efficiency.It is of great significance to guarantee road safety and improve the scientificity of highway maintenance decision.At present,the road condition evaluation method is mainly manual method.Existing methods for potholes identification and evaluation rely on expensive laser equipment and 3d surface reconstruction with high computational costs,or express the severity of potholes according to the area index,which results in the collection of existing potholes detection data can not be able to analyze comprehensive,true and accurate potholes information.To these reasons,another solution is proposed in this paper,which uses 3d reconstruction data based on machine vision to realize automatic detection and evaluation of potholes' severity.Referring to relevant papers,the principle and structure of 3d reconstruction technology,its own characteristics and parameter characteristics are explained in detail,and the outdoor test is set up.The reasons and solutions of the missing points in the process of detection are analyzed.The calibration method of reconstructed data coordinate system based on PCL is proposed.The pothole index method is applied to verify the accuracy of the method.Then,the influence law of point cloud density and number of pictures on the error of potholes calculation index was analyzed.Finally,MATLAB was used to restore the model information,calculate the parameters of the potholes data,which obtains the width and depth of the potholes through three-dimensional reconstruction.The visual and spatial characteristics of the potholes were used to measure the characteristics of the potholes(depth and volume).The method has been verified in practical potholes detection,and the results show that the method has the advantages of low cost and intuitionistic.