一种改进的SIFT特征点匹配算法  

An Improved SIFT Algorithm for Feature Point Matching

作　　者：王立中[1] 麻硕士[2] 薛河儒[3] 侯振杰[3] 

机构地区：[1]内蒙古农业大学职业技术学院,内蒙古包头014109 [2]内蒙古农业大学机电工程学院,呼和浩特010018 [3]内蒙古农业大学计算机与信息工程学院,呼和浩特010018

出　　处：《内蒙古大学学报（自然科学版）》2009年第5期615-619,共5页Journal of Inner Mongolia University：Natural Science Edition

基　　金：教育部"春晖计划"资助项目(Z2007-1-01010)

摘　　要：提出一种改进的SIFT特征点匹配算法.以提高图像特征点匹配算法效率为目的,研究了SIFT特征点描述子基于欧氏最小距离测度的匹配算法.由于SIFT特征点检测算法检测到的特征点数量较大,且每个特征点描述子都是128维的向量,而基于欧氏最小距离测度的匹配算法要求,待匹配第一幅图像的每个特征点要和待匹配第二幅图像的所有特征点求距离,排序后寻找极值,这导致了算法效率较低.依据光学成像理论和双目视觉理论,由第一幅图像每个特征点的坐标,从行列两个方向缩小第二幅图像待匹配特征点坐标的搜索范围,在保持匹配精度的基础上,提高了算法的效率,算法速度约是原算法速度的2.7倍.An improved SIFT algorithm for feature point matching is proposed. The matching algorithm of SIFT feature point descriptor based on Euclidean minimum distance is studied in order to improve the algorithm efficiency of image feature point matching. A large number of feature points are detected with SIFT algorithm, and every point descriptor is 128 dimensions,and the distances between every feature point in first image and the all feature points in the second image need to be calculated and then sorted to find the extremum, so the algorithm efficiency is low. According to optic imaging theory and the binocular vision theory, the area of feature points matched in the second image can be reduced from both row and column directions based on the feature point in the first image. The improved algorithm accelerates the speed of the algorithm with 2. 7 times for the same precision yielded.

关 键 词：SIFT算法 特征点检测 特征点匹配 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]