A SIFT-Based Image Fusion Method

Aimed at the wrong matching appearance when applying Scale Invariant Feature Transform (SIFT) to image fusion. In this paper, we propose to use the SIFT to find the featured points and matched featured points primarily for image fusion. The double Otsu method and average distance threshold segmentation method are used to remove mismatching points and realize an accurate image registration. Finally, we apply the weight average fusion method to blend source images to obtain the final fused image. The results show that the proposed method can automatically realize high precision in fused images and achieve natural image fusion.