基于Bezier曲线的多边形图像骨架矩的研究

摘要

In this paper, we use the theory of skeleton moments to process polygonal images′ skeletons based on the theories of Hu′s moment invariants and Chen′s boundary moment invariants.We express polygonal skeletons with a few of control points by the first-order and second-order Bezier curves, deduce skeleton moment invariants expressed by control points, which provides a statistical feature extraction method for the classification and recognition of polygon images.%在Hu矩与Chen边界矩方法的基础上,将骨架矩理论融入多边形图像的骨架处理中,通过一次与二次Bezier曲线将多边形骨架用控制点表示,推导了基于控制点表示的骨架矩不变量,为多边形图像的分类与识别提供了一种统计特征提取方法.