摘要
文章提出了一种将谱图理论、特征点的局部特征和概率松弛法相结合的特征点匹配算法。该算法通过谱方法,求出特征点匹配的初始概率;利用特征点的结构特征和灰度特征,求得初始支持度;将初始概率、初始支持度与概率松弛迭代法相结合,获得匹配结果。实验结果表明,该方法能够达到较高的匹配效果。
This paper presents an algorithm of point correspondence in which the spectral theory, partial characteristics of points and the method of probabilistic relaxation are combined. The algorithm gains the original probability of point correspondence by the spectral method firstly. Secondly, partial characteristics are used to gain the original support. Finally the original probability and support are combined with probabilistic relaxation to gain the correspondence results. The experiment results indicate that this algorithm can attain to a better matching effect.
出处
《合肥工业大学学报(自然科学版)》
CAS
CSCD
北大核心
2007年第9期1076-1078,共3页
Journal of Hefei University of Technology:Natural Science
基金
国家自然科学基金资助项目(10601001)
安徽省自然科学基金资助项目(070412065)
安徽省自然科学基金资助项目(050460102)
安徽省高等学校自然科学研究项目(2005KJ005ZD)
安徽大学211工程学术创新团队资助项目
关键词
匹配
局部特征
概率松弛
correspondence
partial characteristic
probabilistic relaxation