计算机应用 ›› 2012, Vol. 32 ›› Issue (11): 3161-3167.DOI: 10.3724/SP.J.1087.2012.03161

• 图形图像处理 • 上一篇    下一篇

基于向量长度的Hausdorff距离图像匹配方法

章薇,陈和平,杨玲贤   

  1. 武汉科技大学 信息科学与工程学院,武汉 430081
  • 收稿日期:2012-05-16 修回日期:2012-06-27 发布日期:2012-11-12 出版日期:2012-11-01
  • 通讯作者: 陈和平
  • 作者简介:章薇(1988-),女,湖北黄冈人,硕士研究生,主要研究方向:图像处理、模式识别与智能系统;陈和平(1956-),男,湖北武汉人,教授,博士生导师,主要研究方向:智能信息处理、软件工程;杨玲贤(1978-),女,浙江绍兴人,讲师,博士研究生,主要研究方向:图像处理、模式识别与智能系统。
  • 基金资助:
    武汉市科技技术局资助项目(201110821236)

Image matching method by using Hausdorff distance based on vector length

ZHANG Wei,CHEN He-ping,YANG Ling-xian   

  1. College of Information Science and Engineering, Wuhan University of Science and Technology, Wuhan Hubei 430081,China
  • Received:2012-05-16 Revised:2012-06-27 Online:2012-11-12 Published:2012-11-01
  • Contact: CHEN He-ping

摘要: 针对采用Hausdorff距离描述两组点集之间相似程度时对孤立点噪声敏感的问题,提出一种基于向量长度的Hausdorff距离图像匹配方法。考虑图像中像素点之间的相互关系,连接图像中一个像素点和其他点构成一个向量长度集合,每个像素点对应一个向量长度集合,进行图像匹配时,计算模板图像中每个像素点的向量长度集合与待匹配图像中像素点向量长度集合的改进Hausdorff距离,从而得到匹配的量化结果。实验表明,新方法处理噪声图像具有较好的鲁棒性。

关键词: 向量长度, Hausdroff距离, 图像匹配, 鲁棒性, 孤立点噪声, 相似度量

Abstract: Hausdorff distance is sensitive to acnodes when using it as the measure to describe the similarity of two point sets. Therefore, a new image matching method based on Hausdorff distance of vector length was proposed. Considering the mutual correlation of pixels in the image, one pixel was connected to the others in one image, a set of vector lengths was composed, and then each pixel corresponded to one vector length set. Then, the modified Hausdorff distance between the vector length set corresponding with each pixel in template image and matching image was computed out. At last, quantified image matching results were obtained. The experiment shows that, the efficiency of the new method to deal with image matching problems in random noisy situations is so remarkable.

Key words: vector length, Hausdroff distance, image matching, robust, acnodes noise, similarity degree measurement

中图分类号: