Research on SURF Feature Matching Algorithm Based on Lab Color Space

Abstract

The alignment of images is an important task in the processing of images, which is also known as registration. Among the feature based registration methods, Speeded Up Robust Features (SURF) has been proposed as a computationally efficient approach. The SURF algorithm not only does not consider color information, but the matching accuracy is also rather low. In this paper, an improved SURF algorithm for feature extraction is proposed. An approach is proposed to use the Lab color space to extract the color information of the color image when extracting the feature points based on SURF, and adds the feature descriptors. In this sense, the feature descriptors of different feature points by proposing method are unique. And this method makes up for the shortcoming that SURF cannot use the color information of color images. This improved method promotes the precision of matching. The results show that the matching accuracy of this algorithm is higher than that of the original SURF when physical interference, rotation, brightness change, and scale change occur.