Fusion of Infrared and Visible Images through a Hybrid Image Decomposition and Sparse Representation

Aiming at the fusion of the infrared and visible images, a novel image fusion framework based on hybrid image decomposition and sparse representation is proposed in this paper. Firstly, the Gaussian and guided filters are used to decompose the source images into the small-scale texture details, large-scale edge and coarse-scale image information. The main infrared features are maintained in the large-scale edge information, which are used to determine the fused weights for the coarse-scale information. The sparse representation based fusion method is adopted for the fusion of the small-scale texture details and large-scale edge information, which makes the final fused image can effectively highlight the infrared targets, while preserving the texture details of the visible images as much as possible. So, the fused image is more consistent with the human visual perception effect. Experimental results show that method is superior to the currently used popular image fusion methods.