A image novel fusion algorithm for infrared and visible is proposed

firstly

the images are transformed at multi-scales and multi-directions using nonsubsampled contourlet transform（ NSCT）

and the transformed low frequency subband is fused by improved energy weighting method

and the bandpass subband coefficients is fused by method combined the maximum coefficients with regional variance weighting

then the fused two subband coefficients is reconstructed by inverse NSCT to get fused image. The fused experimental result of different algorithms are compared

it shows that this algorithm has better fusion effect than other fusion algorithms.