METHOD FOR CREATING HIGH QUALITY COMPUTER MODEL FROM ROUGH MODEL

摘要

This paper discusses a data processing technology which enables easy construction of computer models (CAD models) with high quality surfaces from the shape data of roughly created physical models. The position vectors and unit normal vectors of discrete points on the rough physical model are obtained and are processed with the least square method based on quadratic surface approximation. The discrete points are then classified and grouped on the surfaces constructing the model according to the histogram (frequency distribution) of curvature. The position vectors and unit normal vectors of the points classified on each surface are further adjusted by the least square method based on quadratic surface approximation. Finally, the high quality computer model is constructed by combining the improved surfaces.