SIFT image matching algorithm in view of improved SUSAN operator and affine invariance

To improve the matching efficiency of SIFT algorithm, proposing a SIFT image matching algorithm, which in view of improved SUSAN operator and affine transformation. in view of the improved SUSAN operator to extract feature points, using Gabor filter smooth the image, improving feature point position, direction resolution and noise robustness; in view of affine transformation invariance, eliminating mismatched points. Experiments show that the matching efficiency of the algorithm is significantly improved.