Research On The Algorithms Of 3D Point Cloud Preprocessing And Circle Extraction

Reverse Engineering is one method for digital product development, which is more and more widely used in the measure of big scale accessory and complicated curve surface reverse, and the correlative research is recognized more. Data acquisition and data processing are two key technologies of reverse engineering. Data acquisition is the basis of RE, while data processing is the crucial part, because the results affect the quality of the subsequent model reconstruction directly. At present, processing of data cloud based on points becomes a hot point of Reverse Engineering, which includes data pre-processing feature extraction.This thesis reseachs on different approaches of data abtained, and focuses on the topology construction of the point cloud, data points of clouds extraction and circle extraction in 3D points, which make preparations for subsequent corner extraction, compare and model reconstruction. Furthermore, the thesis presents some creative ideas, which were validated in practice.Data acquisition is prophase of data processing. It introduces principle, methods and affection of 3D discrete data acquisition and types of point cloud,also describes the method which the thesis used in data acquisition.Building up the topology of the point cloud is one the crucial technology of data preprocessing. On analysing the common arithmetic of topology of the point cloud, proposing an improved algorithm for searching k-nearest neighbors of scattered-point cloud data and arithmetic flow, this method develops efficiency and accuracy of building up topology of the point cloud.Edge feature extraction from a point cloud is another necessary step for data processing. The thesis discusses edge feature extraction methods which apply to scan line point clouds and scattered point clouds, edge compositor arithmetic is discribed. The paper uses gridding partition to extract edge with experiment data, which makes preparations for detecting and comparing corner points.Circle is a familiar figure in engineering, which makes it important that extracting circle in 3D data points. Relative to complexing methods of detecting circle in edge,the paper brings forward a new extraction methods in scattered points. The detecting arithmetic of ellipse and circle and the extraction methods of points cloud of bold edge are also expounded in detail. In this way,it is easy to extract circle from experiment data.