A New Robust Similarity Measure Based on Lissajous Curves and its Applications in Remote Sensing Image Registration

Similarity measure is crucial to image registration. It decisively affects not only robustness but also accuracy of the registration algorithm. In this paper, an improved similarity measure denoted by MILF is proposed. It is based on trajectories generated from Lissajous figures. Based on MILF and Integrating with noticeably improvement on feature match process by using the edge information, a new registration approach is constructed. The advantages of this registration algorithm is not only analyzed in theory, but also proved by the experiment.