Research And Implement Of Hole-Repairing Technology For Scatted Points Cloud

In the application of reverse engineering, there are often data loss owing to partial covering of the part or limit of the measuring instrument within the acquired origin points cloud. These holes must be repaired in order to form the integral model of the part during the process of the surface reconstruction.This paper proposes an improved hole-repairing method for the scattered points cloud based on existent research achievements, which first distinguishes the boundary points from the data cloud, then fills the hole using the neighbor information of the boundary. All the research and achievement sums up as follows:(1) We have designed an improved boundary extraction algorithm for points cloud with holes and open area. This algorithm searches K neighbor points and extracts the boundary points by the density distribution of K neighbor points, then estimates the direction of the boundary lines and connects the boundary points to the boundary lines, at last distinguishes the inner boundary of the hole from the outer boundary of the open area.(2) We have designed an improved hole-filling method. In this method, points grows from the hole boundary using local information, and fills the hole with points in the process of boundary growing. And the area of new points is connected with the hole boundary smoothly. This method achieves a good performance in the spatial closed hole repair.Besides, a proto system based on these algorithms is implemented in this thesis, and several typical hole-repairing examples are given as well. The results of these examples indicate the correctness of the hole-repairing method.