Multi-Exposure Fusion Method for Space Target Images Based on Multi-Feature Weights

Abstract

Aiming at the characteristics of low brightness of space target image, low contrast with space background and uneven image exposure, a multi-exposure fusion method of space target image based on multi-feature weight was proposed. In this method, contrast features, adaptive optimal brightness features and salience features are used as the basic weights of image fusion weights, and the final weights of image fusion are calculated from these three weights. At the same time, aiming at the low brightness of space target image, local color correction technology is further used to improve the overall brightness of fused image. The experimental results show that compared with the four commonly used multi-exposure fusion methods, the proposed method can achieve better multi-exposure fusion effect on space target images, and the information entropy, spatial frequency and average gradient indexes are improved by 103.8%, 7.69% and 33.72%, singly. IL-Niqe, PIQE and BRISQUE had an increase of 33.99%, 13.57% and 19.07%, respectively.