摘要
本文介绍了一种新的骨架跟踪器,观察可知,任何线结构图形总是由开线、结点线和边界线中的一种或几种组成,新算法分别处理这三种线型。算法的设计力求尽可能地保持图中各几何元素的构成关系。为了实现这一目的,我们引入了上、下、左、右邻点集的概念,线跟踪的方向是通过这些邻点集的比较来决定的。
The article describes a new skeleton follower. It is dcsinged based on the observation: Any line-structured drawing is composed of one or more types of three line types, i.e., opend, feature and closed lines. The follower deals with these three line types respectively. In order to keep the constitutional relations between the geometrical elements in a drawing as far as possible, the concepts of top, bottom, left and right neighbouring sets are introduced. The line following direction is chosen based on the comparison of these neighbouring sets.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
1994年第3期258-262,共5页
Pattern Recognition and Artificial Intelligence
关键词
骨架跟踪器
光栅图像
计算机图形
Skeleton Foilower, CAD Input, Vectonsation, Polygonal Approximations.
