An Improved Vector Algorithm for Contour
Fu Li · Ceshi jishu xuebao · 2002
In reversing techniques based on tomographic reconstruction, the borders of section images are elements used to reconstruct the model of a part. How to extract the geometrical features from contours accurately is the key technique. An improved vector algorithm is presented, in which those edge points which have curvature extremum are extracted as initial segment points. All the segments are classified into two kinds: linear segment and curved segment. Those points on the linear segment or curved segment will be respectively fitted into line or spline curve and then vector. It is proved this algorithm can not only extract the key points which can represent border shape, but also guarantee the reconstructed model have smooth surfaces.