Infrared/visible image matching algorithm based on NSCT and DAISY

Interest point detection and matching are basic computer vision tasks. This paper uses the Non-Subsampled Contourlet Transform (NSCT) detector combined with a DAISY descriptor to develop a robust interest point matching algorithm for infrared/visible images. The NSCT-based detector is very efficient in detecting relevant image features that have good localization and rich geometric information. Once interest points have been extracted, a fast DAISY descriptor can be computed to represent these points, and finally we match them by comparing their descriptors using the Euclidean Distance (ED) and the RANdom SAmple Consensus (RANSAC). The experiment results illustrate that the proposed algorithm has certain robustness for infrared/visible image feature matching.