Improved energy contrast image fusion based on nonsubsampled contourlet transform

An improved image fusion method is proposed in this paper. Firstly, the source image was decomposed into the lowpass and directional band-pass coefficients by non-subsampled contourlet transform (NSCT) for multi-sensor image fusion of the same scene. The low-pass part uses the fusion method of improved the energy contrast, it fully takes into account the energy contrast of coefficient as well as the neighboring coefficient characteristics. The value of the coefficient in the directional band-pass coefficient is related to its surrounding coefficients, and the directional band-pass coefficient uses the fusion method based on context. Experimental results show that the proposed method can improve the quality of fusion image compared with regional energy maximum and regional variance maximum. The proposed method can not only keep the energy of the image, but can also save the edge information of image.