Geometric Feature Detection of Space Targets Based on Color Space

Abstract

The identification and estimation of space non-cooperative target is a prerequisite for space rendezvous and on-orbit service, most of which have geometric structures such as docking loops, antennas, solar panels, and the rest are covered with multiple layers of material for thermal insulation. In this paper, a color space-based method to detect space noncooperative target geometric feature is proposed. First, we split the 3 channels of the RGB image and convert it into another format in HSV (Hue, Saturation, Value) color space. Second, by comparing the 6-channel images under different lighting conditions, it is determined that the Hue channel is used for fine identification of key features. Finally, experiments demonstrate the effectiveness of our proposed method, which applied pass-through filtering method by Hue Channel to detect geometric features under extreme lighting conditions in space.