基于函数拟合的SIFT误匹配点剔除算法

针对RANSAC算法在剔除SIFT误匹配点实验中存在耗时长、误匹配等问题,提出了基于函数拟合的SIFT误匹配点剔除算法。首先采用SIFT算法进行待匹配图像与匹配图像的匹配工作,对匹配后的待匹配图像关键点采用迭代最小二乘拟合法构建函数模型,然后与匹配图像特征关键点进行拟合,最后计算两者的误差,当误差大于给定阈值时,确认该点为误匹配点,对其进行剔除。实验结果表明,采用基于函数拟合剔除误匹配点在时间上比RANSAC算法平均节省了2 s,正确匹配率提高11.75%,并且较多地保留了正确匹配点。

SIFT mismatching elimination algorithm based on function fitting

The RANSAC algorithm has some problems in eliminating SIFT error matching experiment,such as long time,error matching,etc.,so SIFT mismatching elimination algorithm was proposed based on function fitting.Firstly,SIFT algorithm was use to match the images,and then the iterative least square fitting method was used to construct the function model of the key points to be matched.Secondly,the key points of the matching image feature are fitted.Finally,the errors of the two algorithms are calculated.When the error is greater than the given threshold,the point is a mismatch point,and it is eliminated.The experimental results show that this algorithm can averagely save 2 s in operation time comparing RANSAC algorithm,the correct matching rate increases by 11.75%,and more correct matching points are reserved.