3D Laser Scanning Data Processing And Surface Reconstruction Method

3D laser scanning technology is developing in1990s as a new measurementtechnology．by the way of measuring three-dimensional coordinates and digital photos of theresearch objectives, it uses a high-speed and non-contact laser measurement to obtain fast andefficiently three-dimensional information of large entity or real target．The technology hasbeen widely used in nuclear power plants, protection of cultural relics, archeology,construction, shipping, water conservancy, energy, electricity, transportation, public security,municipal construction and other fields．But there are a variety of errors and noise with theoriginal point cloud whose density is general high．So when we use these data for subsequentprocessing and surface reconstruction, it is necessary to preprocess the original data in orderto ensure the accuracy and speed．On the other hand, surface reconstruction involves manyfields, such as geometric calculations, topological relationship, and the drawing of thethree-dimensional model, data processing and surface reconstruction, so studying methods forprocessing point cloud data has major theoretical and practical significance．To solve theabove problem, this article study methods in the following three aspects．1.De-noising．In the analysis of the existing de-noising algorithm based on scattered pointcloud, the article presents a step-by-step de-noising method based on the least squaresmethod．The noise of the point cloud is classified into singularity and small amplitudenoise．Firstly, we remove the singularity by striking a bounding box connected domain;Secondly we fit the points of K-Nearest Neighbor (KNN) a best approximation plane by theleast squares, by determining the size of the distance of each point to the plane and the setthreshold value to remove small amplitude noise．The method achieve the intended effect byexperiments．2.Simplification．The simplified method based on curvature can better retain the detailfeatures of the reconstructed surface．but it is lower efficiency due to the curvature of eachpoint needs to be calculated and compared．So this paper presents a simplified method basedon the contour points extracted．Firstly, we extract the contour points as the feature points ofthe simplified point cloud, and then finding the minimum curvature of the contour points inthe point cloud data as a dividing line, we divide the curvature of the point cloud data into two parts. To greater than or equal to the minimum curvature part, we retain contour points anddelete other points; to less than the minimum curvature part, we process them according to theprinciple of curvature streamline. This method not only inherits the advantage of thesimplified method based on the curvature which can be better to retain the detail features ofreconstructed surface, but also to some extent reduces the computation and improveefficiency.3.Surface reconstruction．There are five surface reconstruction methods: the parametricsurface reconstruction, implicit surface reconstruction, piecewise linear surface reconstruction,subdivision surface reconstruction and deformation of surface reconstruction. We analyzetheir advantages and disadvantages. The process of surface reconstruction using GeomagicStudio is described in detail, and then we using the software reconstructed the point clouddata.