Abstract: Aiming at the problem of low matching rate of traditional feature matching algorithms, we proposed an improved algorithm based on enhanced image gradient information for scaleinvariant feature transform （SIFT） feature matching algorithm. Firstly, a gradient image was obtained by proper gradient operator. Secondly, the gradient image and the original image were fused with the specific weight, and after normalization, the fused image was blurred by Gauss. Finally, the traditional algorithm was used for feature extraction. Experimental results show that the visual angle and invariability of rotation of the i
mproved algorithm are obviously better than those of the original algorithm, and the matching rate of the images with larger brightness or noise is also slightly improved, which effectively improves the accuracy of the SIFT feature matching algorithm.

Key words: gradient, feature matching; local feature, scale-invariant feature transform (SIFT)