InfoScope: Smart telescope system using robust feature points matching under various sunshine conditions

Abstract

In this paper, we propose a smart telescope system, which we call “InfoScope”. The system exploits feature points matching between a pre-registered photo and a live camera screenshot, in order to overlay the POI(Point of Interest) information onto the live camera view. Feature points matching is robust against rotation and scale changes. However, this method is prone to sunshine condition, which significantly affects ambient luminance, to be specific a photo taken in the morning does not match one taken in the evening. To solve this problem, we developed a feature point description database called Om-niReference. To absorb these effects, feature point descriptors from a photo and a photo taken in the evening are merged and stored in OmniReference. By utilizing OmniReference, POI can be successfully displayed on the live camera view accurately even under variable sunshine conditions. We have analyzed photos taken periodically during the daytime from three observation points. This method significantly improves the matching accuracy throughout the day.