Research On Key Technologies Of Sufaces Reconstruction Based On Scattered Point Cloud Data

In recent years, a triangular mesh model from a given point cloud has been sufficiently researched. However, there is the problem of huge data and low quality of triangle mesh. With the development of coordinate measure system, scattered point cloud data which contain more details of the complex shape object, can be obtained conveniently. So, the research of CAD model reconstruction from scattered point cloud should be carried out urgently.The main key techniques contents of surfaces reconstruction from scattered point cloud data include data pre-processing, such as denoising data,filling holes and simplifying data, segmenting data, producing quadrilateral domains, and fitting B-Spline surfaces. The results are as follows:The normal of digitized points is computed by using'micro-tangent plane'. Based on existing adjustment algorithms, a improved normal adjustment algorithm is put forward and impleme nted, which greatly improves the speed of computation. By referring to some classical fairing approaches, a new algorithm of data point denoising is designed on the basis of the modified normal of points. This method obtains perfect effect with the original features preserved entirely compared with other methods of conventional smoothing.Algorithms of curvature estimation and data simplification of scattered point cloud data are studied. Local surface curvature properties on scattered-point data are estimated using a parametric quadric surface approximation, and the key step is local base surface parameteration method which resolves parameteration of scattered-point data. An algorithm is presented for simplifying the scattered point cloud data based on curvature restriction. This simplifying method can preserve critical shape features and simplifies acutely points in flat area.Repairing holes in scattered-point sampled surfaces is discussed as well. A novel hole- repairing algorithm, which can guarantee the consistency between the holes and the original surface, is put forward. The algorithm can be used to repair non-closed holes in scattered-point sampled surfaces and restores some features of original model.Segmenting shapes from scattered point cloud data is premise of reconstruction model accurately. The segmentation process consists of following steps: coarse segmentation in which surface type of each point is identified, refining segmentation which collects points which belong to one surface and cuts noisy points, compiling partitions in order to get...