Image Fusion Based On Wavelet Transform

Image fusion is such kind of process through which multiple source images are combined to generate a more accurate description of the same scene.The fused image is usually more suitable for human vision perception and machine processing tasks.In recent years,image fusion has a wide range of applications in military,remote sensing or medical diagnosis,etc.Aiming at some problems existed in the traditional region-based image fusion methods,a novel compound region segmentation image fusion method is proposed.First of all,two source images are decomposed by DWT,and the each LL band is utilized to guide the image segmentation in the spatial domain alternately.Then the divided regions are superposed to form compound region image.Finally,image fusion is performed in wavelet domain,referencing to compound region.Compared with the traditional region fusion,the proposed algorithm overcomes the problems,such as inaccurate region segmentation and poor fused result for two images with different resolution.The last experiment results show that,the novel fusion rule is capable of getting better fusion effect,whatever in subjective or objective evaluation standard.