Comparative Analysis of IR-IR Image Matching Using Classical Techniques

Abstract

In this paper, after finding feature points and description using classical feature detection and description techniques, this paper compares and analyzes the performance of each classical technique by matching descriptors of the IR-IR images with the FLANN technique and matching with the Homography technique. Usually, these classical techniques are often used in EO images, but recently, in technologies that require a high level of detection such as defense technology, not only EO images but also IR images are being used. Therefore, classical feature detection and description techniques specialized for reflected light are applied to IR-IR images, and their performance is compared, and the limitations of classical techniques are examined. Classical techniques to use include SIFT, SURF, and KAZE.