基于势能平衡的图像骨架抽取算法

Potential balance-based image skeleton extraction algorithm

基于骨架的基本特征和势能的概念 ,提出了一种基于势能平衡方式的二值图像骨架抽取算法 .将抽取对象的边缘点作为零势能的等势点 ,计算出处于势能平衡点上的基本骨架点 ,并据此通过向势能小处延伸的方式构建出骨架 ,还解决了由于各点周围参考零势能点数量不同而带来的各点势能值间难以比较的问题 .通过计算机模拟显示 ,本算法对于边界噪声影响不太敏感 。

Based on the basic features of skeleton and the concept of potential, a 2-level image skeleton extraction algorithm based on the balance of potential is proposed in this paper. By treating the edge pixels of image object as 0-potential points, basic skeleton points which locate at the potential-balance points are discovered, and a skeleton is formed later by extension. At the same time, this paper deals with the difficulty of comparing potential value between different points, which is caused by the different numbers of reference points. The simulation results show that the algorithm is insensitive to the boundary noise, and distortion at the end point is less than that of the algorithm based on the biggest disc.