采用自适应搜索范围的多介质立体匹配算法  

A multi-medium stereo matching algorithm using adaptive searching area

作　　者：李海滨[1,2] 刘婷婷[1] 张强[1] 张文明[1,2] 李雅倩[1] 

机构地区：[1]燕山大学工业计算机控制工程河北省重点实验室,河北秦皇岛066004 [2]燕山大学国家冷轧板带装备及工艺工程技术研究中心,河北秦皇岛066004

出　　处：《中国科技论文》2014年第4期456-459,464,共5页China Sciencepaper

基　　金：河北省自然科学基金资助项目(D2014203153;D2013203363);高等学校博士学科点专项科研基金资助项目(20101333120006)

摘　　要：摄像机在水下拍摄时,成像光线经过不同介质会发生折射,空气中极线约束条件不再满足,区域匹配无法在此极线上搜索。针对此问题,提出一种采用自适应搜索范围的多介质立体匹配算法。首先对双目摄像机进行标定得到相关参数;然后针对每一像素点推导出曲线极线,并计算出与原极线的上偏差与下偏差,确定该点的最优搜索范围;再用区域算法SAD算法在每点对应的最优搜索范围中进行多行搜索并匹配。经过两组实验对比,该算法的匹配精度分别为72.4%、76.8%,而SIFT匹配算法的匹配精度分别为71.1%、74.7%。该算法的匹配精度达到了特征匹配的精度,可以用于整幅图的稠密匹配,解决了原极线搜索策略无法应用于多介质环境的问题,成功将区域匹配算法应用于多介质中的立体匹配。Light refraction occurs when imaging in the water,the epipolar constraint of air is no longer valid,and thus region matching method cannot search on this epipolar.This paper presents a multi-medium region matching algorithm based on adaptive searching area.First,the binocular camera should be calibrated in order to obtain some relevant parameters;then the upper and lower deviations from the line in the air can be calculated through the curve constraint and the optimum searching area is therefore determined.Next,the SAD region matching algorithm can help to match two images,at the same time,instead of searching on the original epipolar line,we propose an adaptive searching area so that the searching is performed in this area with several lines to achieve the purpose of a higher accuracy.The test results indicate that our algorithm achieves the matching accuracy of 72.4%and 76.8%,while the matching accuracy of the SIFT feature matching algorithm is 71.1% and 74.7%.It suggests that our algo-rithm achieves the same matching accuracy as the SIFT feature matching algorithm and this can be used to perform dense matc-hing.The problem that the existing searching method cannot be used in multi-medium is therefore solved.And the region matc-hing algorithm is successfully applied to multi-medium image matching.

关 键 词：机器视觉 自适应搜索范围 多介质 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]