An Enhanced Feature Matching Method for Multi-modal Remote Sensing Image Registration based on Multi-dimension Constraint

Abstract Multi-modal remote sensing image registration is the key foundation of remote sensing image processing, which is also a significant research topic in the fields of environmental modeling and Earth detection. The characteristics of multi-modal images, such as variations in radiation, geometry, scale, viewpoint, and dimensionality present significant challenges for achieving high-precision matching. Aiming at increasing the registration points when the error is similar, this paper proposes an enhanced feature matching (EFM) method for multi-modal remote sensing images, which includes: 1) An low-complexity moment (LCM) calculation for a modified feature point extraction method; 2) Multi-dimensional space constraints (MSC) joint of phase, position and direction. The experimental results show that the EFM method has achieved significant improvement in feature point extraction and matching of multi-modal remote sensing images, with a three fold increase in registration points compared to conventional registration schemes, making it suitable for remote sensing image registration.