Image Fusion Based On Region Energy

Image fusion based on region energy is a hotspot researched recently on image fusion area. The relativity between pixels in a image is calculated in image fusion based on region energy, local character of a image got even more materialized. Image fusion based on region energy overcomes the shortcoming that the relativity between pixels in a image don't get calculated, which is in the image fusion based on pixel rule. The fusion quality got enhanced.Wavelet transform is the base of mathematics in this paper. Three parts are basically discoursed.Firstly, The traditional image fusion based on region energy got improved on. Matching degree calculated by use of the traditional image fusion rule based on region-energy maybe is negative because of the incomplete registration among different images, which results in the difficulty of estimating the difference between two different regions from the viewpoint of energy. The traditional matching degree is modified by the introduction of absolute value, and a new fusion rule based on such a modification is proposal in this paper.Secondly, this paper aims at the image rule of pixel and the image rule of region energy, different wavelet bases are chosen, entropy, average value, standard error, visibility and mean-cross entropy are used to evaluate the result of the experiments. Thirdly, During the method of image fusion in use of wavelet decomposition and construction , weighted average was almost selected in the rule of low frequency, but it will be a question how to select reasonable coefficients. So in this part a auto-adapted method of image fusion based on region energy was proposed, which would select coefficients auto-adapted, which solved the question simultaneously.