Surface 3D Reconstruction And Texture Depth Index Extraction Based On Laser 3D Scanning

The rapid development of highway construction in China puts forward higher requirements for modern maintenance management.Timely and reasonable maintenance decision-making requires real-time and accurate grasp of various technical indicators of the road.The value of the macroscopic texture depth of the road is one of the important indicators for evaluating the anti-skid performance,which affects the driving comfort,fuel consumption and noise pollution.Traditional manual or semi-automatic detection methods are greatly affected by humans’ manipulation.The current road multi-function detection equipment has problems of environmental interference,data interruption,and high cost.The application of three-dimensional scanning technology to road detection has a good prospect.Therefore,this article is committed to developing a high-precision,anti-interference,moderate cost,and intuitive display of the road surface 3D detection equipment and feature extraction technology,which lays the foundation for the subsequent detection of other indicators and the formation of a complete evaluation system.Firstly,this thesis discusses the characteristics and application scenarios of the current common machine 3D vision technology,analyzes the needs of road detection and selects the principle of laser triangulation as the detection technology.On the basis of the current industrial inspection equipment,discuss the equipment scheme applied to the road surface,select the hardware and formulate the inspection process of the scanning system.Secondly,this thesis conducted a simplified mathematical model of the system based on the analysis of the scanning principle of the system and the conversion relationship between camera imaging,and the unknown parameters are calibrated to obtain the mathematical expression of the model by using Zhang’s calibration method and the principle of constant cross-ratio.A special gauge block is designed to verify the accuracy of the system after calibration.Then,this paper proposes an improved Steger method to extract the center point of the light strip image.The ROI area of the light strip is automatically segmented by threshold detection to reduce redundant calculations,and the smooth function is combined to fit the center line to remove the discrete noise on both sides.Compared with the gray barycentric method,the calculation efficiency is reduced,but the algorithm has high precision,anti-interference and strong universality.In the extracted three-dimensional point cloud data,through the cubic spline interpolation fitting of the contour data,the data is distributed at equal intervals,part of the missing data is supplemented,and converted into a depth matrix,which greatly improves the extraction and storage efficiency of the system,and realizes the three-dimensional reconstruction of the road surface.Finally,this paper uses the three-dimensional data of the pavement to propose an MTD calculation method based on the MTD calculation datum.The Monte Carlo expectation method is used to calculate the volume of the envelope between the datum and the road to calculate the MTD value.The comparison results of the on-site sand patch test and the scanning test are highly correlated,which shows that the system has high scanning accuracy and the calculation algorithm is stable and effective.In this study,a laser three-dimensional scanning equipment of pavement surface was developed,which realized the extraction of three-dimensional data of road surface and matrix conversion,and proposed the corresponding MTD calculation algorithm,which has high engineering practical value.