一种二值图像连通区域标记的新方法

A New Algorithm for Binary Connected Components Labeling

论文提出了一种基于区域生长的二值图像连通区域标记的快速算法。与传统方法相比,该方法的特点是在一次图像扫描中完成所有连通区域的标记,而且避免了大多数改进算法都必须处理的重复标记的问题;同时,该方法不受所标记的图形形状的影响,表现出良好的算法鲁棒性。最后分析了算法的计算复杂度,并与传统算法和两组改进算法进行了比较,试验结果表明了算法的高效率和鲁棒性。

A new algorithm based on region-growing for binary connected components labeling is proposed in this paper.Compared with the traditional method,the characteristic of the algorithm is that all the connected components are labeled in a single scan of the input binary image and the problem of labeling redundancy,which most improved methods have to deal with ,is avoided.Otherwise ,the computation of the algorithm is independent of the shapes of the connected components.At the end of this paper,the computation complexity of the algorithm is discussed,and comparative experiments with traditional method and two of its improved methods are made.The experimental results show the proposed algorithm high efficiency and robustness.