一种改进的边界轮廓矢量化算法  

An Improved Vector Algorithm for Contour

作　　者：付丽琴[1] 陈树越[1] 

机构地区：[1]华北工学院电子信息工程系,山西太原030051

出　　处：《测试技术学报》2002年第4期241-244,共4页Journal of Test and Measurement Technology

摘　　要：在基于断层测量的反向建模中 ,物体断面图像序列的区域边界是实现模型重构的基础 ,如何从边界轮廓精确提取几何特征是实现重构的关键 .本文提出一种改进的矢量化算法 ,该算法的基本思想是首先提取曲率极值点作为初始分段点 ,然后判定每两个相邻初始分段点之间轮廓线的形状 ,并区别直线段与曲线段分别进行拟合和矢量化 .实践表明 ,该算法既能准确提取边界形状的特征关键点 。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.

关 键 词：断层测量 矢量化 轮廓跟踪 曲线拟合 曲率极值点 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] O18[自动化与计算机技术—计算机科学与技术]