一种简化的SIFT特征匹配算法研究

Research on the Simplified SIFT Feature Matching Algorithm

针对经典SIFT算法因特征描述符维数过高而导致匹配效率降低的问题,本文提出了一种简化的SIFT特征匹配算法,首先对算子进行降维,以提高速度,然后用双向匹配消除错配以保证算法的准确性,实验取得了较好的结果,验证了方法的可行性。

For SIFT algorithm,for the matching problem of the classic descriptors for feature dimension is too high and lead to reduced efficiency,this paper presents a simplified SIFT feature matching algorithm,the first of the operator dimensionality reduction to improve the speed,then use two-way matching to eliminate errors together with the algorithm to ensure the accuracy of experiments and achieved good results,verify the feasibility of the method.