计算机与数字工程Issue(12)：2260-2265,6.DOI:10.3969/j.issn.1672-9722.2015.12.035
基于视差估计的图像双向匹配
Bidirectional Image Matching Method Based on Disparity Estimation Indexing
邓坚瑾 1田金文1
作者信息
- 1. 华中科技大学自动化学院 武汉 430074
- 折叠
摘要
Abstract
A novel method is proposed in this paper to extract key points from images .It performs fast ,accurately and can obtain dense matching .Firstly ,bidirectional method is used to conduct image matching through sparse key points extrac‐ted from two images .Then ,using correlation coefficient as the similarity criteria ,cooresponding points of dense key points will be searched around small regions estimated by means of the sparse matching in the first step .Finally ,part of those matching pairs with abnormal disparities will be removed to obtain a fast and accurate matching .Experimental results indi‐cate that the proposed method performs well on image matching rapidly and precisely .The proposed matching method to search cooresponding points of prime key points among sub prime key points avoids missing cooresponding points of prime key points .Bidirectional matching method is used to obtain more reliable results .A searching method among small regions through disparity estimation indexing makes the matching process much more quickly ;The proposed constraint criteria based on the standard deviation of disparity can eliminate mismatching effectively .关键词
图像匹配/视差估计/最大梯度/主兴趣点/次兴趣点Key words
image matching/disparity estimation/maximum gradient/prime key point/sub prime key point分类
信息技术与安全科学引用本文复制引用
邓坚瑾,田金文..基于视差估计的图像双向匹配[J].计算机与数字工程,2015,(12):2260-2265,6.
