摘要
提出了一种在核线几何约束下基于自适应归一化互相关(NCC)及奇异值分解(SVD)的尺度不变特征变换(SIFT)特征匹配算法(SVD-NCC).算法首先利用SIFT特征的尺度和方位信息对特征点邻域进行仿射变形改正,然后基于NCC测度和SVD算法生成特征点间匹配矩阵并获得其对应关系.在具体实现策略上,算法首先基于特征点的空间分布和信息熵选取一定数量的最优SIFT特征点集,并基于SVD-NCC算法获得初始匹配点对用于立体像对的核线几何估计,然后在核线几何约束下对其余特征点进行扩展匹配及误匹配剔除.实际的宽基线序列立体影像匹配试验结果表明该方法可显著提高匹配点的数量和匹配正确率.
A new feature matching algorithm based on the singular value decomposition(SVD) was proposed(SVD-NCC).In this algorithm,affine deformation was firstly corrected for matching window by using the scale and orientation information of SIFT features,and then the matching matrix was established base on the normalized cross correlation(NCC) and SVD,and the correspondences can be determined based on the matching matrix.In the practical strategy for this algorithm,the optimal SIFT features with good spatial distribution and large information content were first selected,then these SIFT features were matched by using the SVD-NCC algorithm,and then the fundamental matrix can be estimated by using these initial correspondences.Other SIFT features were matched by using epipolar geometric constraint.The test results for the wide baseline image sequences indicate that the proposed algorithm can increase the amount and accuracy of corresponding points.
出处
《中国矿业大学学报》
EI
CAS
CSCD
北大核心
2011年第6期964-968,共5页
Journal of China University of Mining & Technology
基金
国家自然科学基金项目(41001312
41001297
40901221)
中国博士后科学基金项目(20090450182)
关键词
尺度不变特征变换
归一化互相关
奇异值分解
核线几何
scale invariant feature transformation
normalized cross correlation
singular value decomposition
epipolar geomet