Research On Tin Generation Algorithm Based On Pavement Point Cloud

Detection of road maintenance is now mainly rely on manual to complete, The method of manual testing is inefficient, At the same time subject to weather and other environmental factors. Laser scanning technology is one of the rapid development of technology in recent years, It can quickly and accurately obtain information on the spatial coordinates of the object. At the same time three-dimensional modeling studies related algorithms makes the establishment of a three-dimensional model of the road becomes possible. This article is in this context of research related algorithms for three-dimensional surface modeling technology and the establishment of a three-dimensional surface modeling system.As used herein, laser scanners and gamma camera to obtain the point cloud and image information of road pavement. The acquired data passed into the computer, After a series of processing algorithm to obtain a three-dimensional model of the road.Firstly the preprocessing algorithm for point cloud data, including point cloud point cloud data patching algorithm and simplified algorithms. Point cloud data is to repair the missing point cloud data reasonably assigned according to road conditions, Combined with road conditions for the missing part of the repair area to take two different ways, According to repair cracks and repair in accordance with the flat surface, For most of the pavement is missing the point cloud, the algorithm can correct repair missing the point cloud data. For point cloud simplification, reference angle- chord algorithm, based on a simplified algorithm slope difference of point cloud, Under the premise to retain Pavement point cloud feature points can be simplified out 90% of the point cloud data. In order to get a more realistic three-dimensional surface model, using irregular grid algorithm, using Delaunay triangulation based on divide and conquer algorithm processing point cloud data to generate three-dimensional model, In order to improve the efficiency of generating three-dimensional model, Use parallel computing to improve the efficiency of the Delaunay triangulation divide and conquer algorithm. This parallel Delaunay triangulation algorithm based on divide and conquer can establish the right three-dimensional model of the road in a shorter time. In order to quickly locate surface features, This paper using the method according to the average height of the road and triangle normals of the road surface to identify and locate the road characteristics. Both algorithms can correctly identify the location and characteristics of the road surface feature location. Finally, the establishment of a three-dimensional surface modeling system based on the above theoretical research, Including online testing system and Offline roaming system. The system uses a C ++ implementation of each module functions. Using Open GL, multithreading, XML file parsing techniques in this system.Experimental results show that the established online testing system can collect real-time road information, identifying road features, and fast 3D model of the road for the user to observe. Offline roaming system can generating three-dimensional model based on user settings, and users need to roam the entire road surface model to complete the testing work on the road.