Image fusion algorithm for visible and PMMW images based on EM and Ncut

Aiming at the fusion of visible and Passive Millimeter Wave (PMMW) images, a novel region-based algorithm based on Expectation-Maximization (EM) algorithm and Normalized cut (Ncut) algorithm is proposed. Firstly, it takes advantage of the particular ability of PMMW image in presenting metal target or other plastic bomb and uses the region-growing for PMMW image to extract the potential target regions, then visual and PMMW images are weighted to get a composite image according to the contrast of potential target regions in PMMW image and background around the potential targets in visual image. Next, Ncut is applied to the composite image to obtain the appropriate region mapping image to be modeled. At last, a statistical image formation model with Gaussian mixture distortion is built and EM is used region-by-region to produce the fusion image. Since we model potential target regions within a region, this algorithm could be more meaningful than pixel-based methods and save computation cost. Meanwhile, it is excellent in suppressing noise. The experiments demonstrate the superiority of the proposed algorithm for metal target detection compared to wavelet, contourlet method and rectangle-based EM algorithm.