基于游程连通分析的轮廓提取算法

Boundary extraction algorithm by means of run-length connectivity analysis

通过设计一种称为子轮廓的数据结构,对游程左右端点以及分叉点和交汇点进行有序链接,提出一种新的基于游程连通分析的轮廓提取算法.按照从上到下、从左到右的顺序扫描图像数据进行游程编码,根据当前行游程与相邻上一行游程间的连通情况进行子轮廓创建、子轮廓生长以及子轮廓联结三种操作,实现图像内外轮廓以及内外轮廓包含关系的提取.实验结果表明,算法在提取游程平均压缩率高的图像轮廓时具有较大的优势.

A sub-contour data structure was designed to link the left and right endpoints of runs and the splitting and merging points in order. On the basis of run-length connectivity analysis,a new boundary extraction algorithm was proposed.The run-length encoding was implemented by the algorithm through a successive scan.The connectivity between runs in current and the above rows was analyzed.Three operations,including sub-contour creation,sub-contour growing and sub-contour linking,would occur in different connectivity situations to implement generating outer and inner contours and contours inclusion relationship.Experimental results show that this algorithm owns higher performance than that of other algorithms for the images with high run average compression ratio.