一种基于边缘点特征的细化算法

A Thinning Algorithm Based upon Edge Pixel Feature

本文根据二值图像的边缘点在8个方向上具有较多的像素不相互对称这一重要特征,提出了一种新的细化算法.由剥蚀与矫正两部分组成.在剥蚀部分,只需判别被测像点的8个或12个邻点的分布情况,就可以决定该点删除与否.矫正部分则是对剥蚀后的图像的进一步完善.本文所介绍的算法具有判别简单、对噪声不敏感、并能很好地保持原模式的形状及连通性的优点.

This paper describes a new thinning algorithm according to the important feature that there are more unsymmetrical pixels in the edge points of black-white picture in eight directions. This algorithm is composed of two steps,i.e.the stripping and the rectifying.In the stripping step,it is possible to determine whether a point will be deleted or not only by judging the distribution situation of eight or twelve adjacent points of the point determined.The quality of picture which has been stripped is improved by the rectifying step.The algorithm mentioned here is simpler in determination,insensitive to noise and main- tains well the shape and the joint of former pattern.