Image registration is a well researched topic of computer vision. To deal with matching efficiency, repetitive pattern matching and affine invariant matching better, two improvements over the state-of-the-art Affine-Scale Invariant Feature Transform (ASIFT) algorithm were presented. The feature extraction of matching frame was developed to improve the matching efficiency of the ASIFT algorithm. The second increased the accuracy of matching and the adaptive capacity of repetitive patterns through the use of improved matching algorithm by combining Optimized Random Sample Consensus (ORSA) with Random Sample Consensus (RANSAC) algorithm based on geometric linear constraint model with homography matrix. The experimental results show that the proposed method is able to well match highly repetitive patterns and has smaller calculation, faster speed and higher accuracy as well.