二值图像连通域标记优化算法

Optimizing Algorithm of Labeling Connected Components in Binary Images

在分析现有二值图像像素扫描连通域标记算法的基础上,提出像素扫描连通域标记优化算法。本算法具有速度快,算法简单,易于实现的特点,仅需两次扫描,即可实现像素的多目标标记。本算法将背景也作为目标加以标记,分两步扫描图像和临时连通域标号矩阵完成连通域的标记和合并,采用顺序存储结构存储和处理等价标号,算法速度快,节约内存。

Based on analysis of existing algorithms of labeling connected components through scanning pixels in binary images,the improved algorithm is proposed.The algorithm has more advantages such as faster,simpler and easier to realize than the others,only need two-scans to label multi-objects which pixel belong to.The algorithm could label background as objects,labels and merges connected components via the two procedure of scanning binary image and scanning the matrix of provisional labels.It adopts sequential storage structure to store and resolve the equivalent labels,runs more rapidly,occupies much less memory than the unimproved algorithm.