An image fusion algorithm based on iterative wavelet transform for space non-cooperative targets

Abstract

Due to the interference of outer space environment, the real image data of non-cooperative targets collected by a single sensor often has problems such as low contrast and insufficient color information display. In this paper, we propose an image fusion method with iterative wavelet transform, which utilizes the respective advantages of visible image and near-infrared image to improve the clarity of space non-cooperative target and retain color information. The dataset in this paper comes from the satellite model group taken by ourselves. Due to different shooting angles and resolutions, we need to carry out accurate registration of near-infrared images and visible light. Therefore, we also propose an improved registration algorithm based on ORB feature matching. In order to verify the effectiveness of the method, other image fusion methods are selected for comparative analysis. The method proposed can preserve texture features and color information of space non-cooperative targets.