Color fusion based on EM algorithm for IR and visible image

As two of the most commonly used image sensors in night-vision for military applications, infrared (IR) and visible light cameras have their own advantages and limitations due to the physical differences between the image formation processes. In this paper, a new multi-scale scheme for color image fusion is presented using EM algorithm and color transfer algorithm for IR and visible image. The color fusion algorithm implements in YUV color space. The Y component is obtained from the result of IR and visible image fusion, by which the original images are fused in nonsubsampled contourlet transformation by using EM algorithm. U and V components come from the original images by simple arithmetical operation. The V component is enhanced by a ratio of local to global divergence of the IR image. The experiment results demonstrate by real image data that the proposed scheme can achieve better fusion performance than the traditional color image fusion methods.