Multi-resolution Wavelet-based Pixel-level Image Fusion Research

Image Fusion is combining images or image sequence signals that get from two or more sensors at the same time (or different time) into a new single image. This new image can't get from any single senor. The purpose of image fusion is to reduce uncertainty. Image fusion plays an important role in the fields of medicine, measuring, geographic information system, industry, intelligent robot and military affairs.The article first introduces the history and some theories of image fusion, gives some quality assessments of image fusion, and then introduces both algorithms of traditional image fusion and multi-resolution wavelet-based image fusion, and put emphases on Mallat-based image fusion. After lots of studies of multi-resolution wavelet-based image fusion, this article put forward a new algorithm of multi-resolution wavelet transform image fusion, the effect of this algorithm is all the better, the fusion image is more clear.The algorithm in this article is using Mallat multi-resolution wavelet transform to decompose and reconstruction images, base on some image pixel position, judge by the area energy and the area spatial frequency of this pixel, we can confirm the pixel position in the fusion image. This algorithm not only effectively reserves the useful information, but also effectively removes the redundant information. The edge and detail of the fused image that gets form this algorithm is more clarity.