Infrared and visible image fusion based on local gradient constraints

Infrared and visible image fusion is the technology that aims at fusing thermal radiation information and texture information, which are respectively contained in the infrared image and visible image, to obtain a satisfactory quality fusion image. In this paper, a new variation model on the basis of local gradient constraint is proposed for texture enhancement. Unlike other methods that only focus on the global constraints to extract texture information in visible image, we focus on the local gradient change to extract detailed texture feature. Experiments compared with SOTA algorithms show the efficiency and priority of our method.