Research on Ghost-Free High Dynamic Range Image Fusion Technology of Space Non-cooperative Target

Failed satellites with large nutation and precession are usually in a tumbling state, in order to achieve emergency rescue of failed satellites, the service spacecraft needs to have the ability to measure the relative pose and attitude of specific parts of the ultra-close-range tumbling target under complex lighting conditions. However, most of the satellite surfaces are covered with thermal control materials with specular properties, the images obtained by the visual sensor under complex spatial illumination are overexposed or too dark, which makes it difficult to extract target features. A method based on motion map detection and changing the weight map of multi-exposure fusion is proposed to implement the operation of eliminating ghosting. The proposed method can solve the problem of weak dynamic range in the image and the ghosting problem in the process of fusion of multiple images. Even when there are two images with a wide exposure range as input images, the above problem can be solved. Compared with the prior art, the method can obviously improve the visual image and reduce the ghost phenomenon. Furthermore, our method has lower computational cost and complexity, especially for mobile phone related applications, which makes the implementation more efficient.