Abstract: The dual-tree complex wavelet transform is employed in image fusion, to round the deficiencies of existing wavelet based algorithms in characteristics representation. The study of Robins et al.. shows that the local energy is robust in the representing and locating of all kinds of image features. Based on dual-tree complex wavelet transform, the local orientation energy and local energy are defined, which in conjunction with the response characteristic of human visual system to image features, the local banded contrast is finally formulated. In the real time image fusion system, inputs may be corrupted with random noises. The local orientation energy entropy is formulated, according to the different orientation energy distribution of features and noises, to modulate the local banded contrast adaptively. As a result, the robustness to noise is improved. The performance evaluations of the proposed algorithm both subjectively and objectively manifest the excellent efficacy of the new scheme.