Novel image fusion algorithm with novel performance evaluation method

By revising the definition of the cross entropy and combining it with entropy, a novel image fusion performance evaluation function based on entropy and cross entropy, i.e., the associated entropy is presented. By use of the associated entropy as the objective function, a novel image fusion algorithm is proposed, which is manipulated by operating the maximum method and the weighted average method alternatively and iteratively. Experiments are done to multiple groups of images, including the fusion of inferred image with optical image, and the fusion of SAR with optical image, and the fusion of optical images with different properties etc. The results concluded from the experiments verify the feasibility and effectiveness of the novel image fusion method as well as the novel performance evaluation function.