Image Registration Based on Corner Cauchy-Schwarz Divergence

A new matching metric-Cauchy-Schwarz divergence of corners (Corner Cauchy-Schwarz Divergence, CCSD) is defined, and a new approach for image registration based on CCSD and sub-pixel localization is proposed. Firstly corners in an image are detected by multi-scale Harris operator, which are taken as initial interest points. And then sub-pixel localization technique is applied to determine the location of corners and eliminate the false and unstable corners. After that, CCSD of all corners is calculated, so matching corners can be obtained. Finally, RANSAC is used to estimate the parameters robustly based on initial matching. The experimental results illustrate the algorithm has a good performance of accuracy and efficiency.