基于对称点对序列的带状图骨架化算法研究

A Skeletonization Algorithm Based on Symmetric Point Pairs Sequences of the Ribbon Image

文中提出了一种基于离散轮廓的对称点对序列的骨架化算法。首先提出带状图像的对称点对序列的概念,它是通过轮廓离散点的Delaunay三角剖分而获得的,并将这个序列应用于带状图的骨架化算法中。在带状图的交叉区域利用三角剖分对偶图模型对该序列进行归并重构,从而得到交叉区域不畸变的骨架。实验表明,该算法充分利用了带状图的全局和局部信息,产生的骨架准确地反映了带状图像间的拓扑关系。

In this paper,a skeletonization algorithm based on Symmetric Point Pairs Sequences（SPPS） of the ribbon image is described.Firstly,the conception of SPPS is proposed.It is obtained through the Delaunay Triangulation （DT） for sampling points of the contour image.Later,this conception is involved in the skeletionation for the ribbon image.In the intersection region of the ribbon image whose structure is represented by the dual graph model of triangulation,the SPPS is merged and reconstructed.So the skeleton that isn＇t deformed is obtained.The experimental result shows that this method make full use of the global and local information of the ribbon image and the skeleton preserved the precise topological relation between the ribbon images.