Image Fusion Based on Multi-objective Optimization

This paper solves the image fusion problem by using a multi-object optimization strategy and a key energy function. The energy function mainly consists of two components. One ensures injection of more correlated detailed spatial information. The detailed information is extracted from gradient representation of the image to be fused. The other one guarantees that the spectral information is preserved by a data fitting term. By minimizing the proposed energy function, the fusion result can be obtained. Moreover, a key parameter is used in the energy function to adjust the weights of the spectral and spatial information during the image fusion. In this paper, a multi-object optimization is constructed to determine such a key parameter. The image fusion performance is evaluated through visional perception and some fusion indexes. Experimental results further demonstrate advantages of the proposed technique over the conventional fusion techniques.