Adaptive Scene-Matching Algorithm based on Frequency Pattern Analysis for Aerial Vehicle

Abstract

Vision-based navigation can be classified primarily into relative and absolute navigation. Absolute navigation can estimate stable position solution of the vehicle. However, the performance of the feature detection algorithm greatly varies depending on the environment, and the matching accuracy may not be guaranteed. This represents navigation accuracy and can lead to serious navigation errors if an appropriate feature detector is not used. Therefore, in this study, we propose an algorithm that uses optimized feature information detector through frequency analysis of input image. To verify the performance of the proposed algorithm, we used the data obtained from flight experiments. The proposed algorithm prevents the occurrence of mismatching in flight experiments.