Image fusion using MGA based on sub pixel

Split the pixels in rectangular or circular region into subpixels. Each subpixel has a weight in inverse proportion to the distance between it and the center pixel. All subpixel weights in a pixel make up the compound pixel weight. In this paper, we obtain image decomposition tree using contourlet, then select average of the coarse parts, and select the maximal parts of the detail components by using region weighted energy fusion rules based on subpixel. Our experiments show that this algorithm has better fusion performance than the image fusion schemes base on traditional region energy. From the point of view of information entropy and image definition, it can improve the quality of fusion image.