Comparative Analysis of Mainstream Multi-Source Image Matching Methods on Remote Sensing Imagery

Accurate multi-source remote sensing image matching is the foundation of many applications. However, the complex geometric and nonlinear intensity distortion between remote sensing images makes it challenging. This work selects 11 representative ones: OS-SIFT, PSO-SIFT, LGHD, LINIFT, RIFT, NISR, POS-GIFT, RedFeat, MatchFormer, and SemLA, and compares their performance on various remote sensing images. At last, we summarize the work by giving the research bottleneck and potential research directions toward high-precision multisource remote image matching.