Infrared and Visible Image Fusion Based on TRPCA and Visual Saliency Detection

The aim of infrared (IR) and visible (VI) image fusion algorithm is to produce an informative fused image which integrates the bright targets features from the IR images and original visual details in the VI images. This paper proposes an IR and VI image fusion method based on tensor robust principal component analysis (TRPCA) and visual saliency detection. With TRPCA model, the source images are decomposed into low-rank part and sparse part. Thus, the details and significant targets can be effectively separated from the source images. The low-rank parts are fused using contrast visual saliency detection, which can reduce the contrast degradation of the fused image. The sparse parts are fused using the local energy fusion rule, which can effectively fuse the significant target information in IR image. Experiments show that the proposed method is significantly better than seven state-of-the-art fusion methods in terms of both subjective visual assessment and objective indexes. Extensive experiments are conducted on 8 pairs of IR and VI source images. 7 state-of-the-art fusion methods are also used to compare with the proposed method. The experimental results show that the proposed method is superior to other methods in terms of both subjective and objective evaluations.