Robust image matching based on the information of SIFT

Abstract

Image matching is an important problem in computer vision. However, due to large geometrical differences and appearance changes in acquired image pair, common image matching algorithms often have a large number of incorrect matching point pairs and the matching accuracy is low. In this paper, we propose a robust image matching algorithm based on the combination of the wavelet transform and Scale-invariant feature transform (SIFT). Firstly, we adopt Discrete wavelet transform respectively on a reference image and a template image to extract their low frequency parts, then we use harris corner detection to detect and match the interesting points in their low frequency parts to determine the matching candidate region of template image in reference image. Furthermore, we extract SIFT features on the matching candidate region and template image and match them by k-d tree and bidirectional matching strategy. Finally, we exploit the information from SIFT to comprise matching constraint and use them to get more correct matches. Experimental results show that, the algorithm can improve the accuracy of matching.