Research And Application Of Image Fusion Algorithm Based On Multi-scale Decomposition

Image fusion is to extract the advantageous information in greatest extent of the same target from different channel through the complementary and redundancy, to merge into a high quality image, and has been applied in various fields. This paper studies the image fusion algorithm based on multi-scale decomposition. The main research content is as follows:1) Proposes the adaptive image fusion algorithm based on wavelet transform(DWT).Combined the strategy based on the window with weighed average method. Puts forward the fusion rule of the low frequency coefficients based on the local region-energy ratio, and the high frequency used an adaptive algorithm based on the local region-energy matching. And at the same high frequency fusion rule, compares with two kinds of fusion algorithm which based on the weighted average and the choose-max strategy based on regional energy. The experimental results verified the effectiveness of the proposed algorithm.2) Studies an image fusion algorithm based on dual-tree complex wavelet transform(DT-CWT), puts forward different fusion rules according to different characteristics of high-low frequency. The low frequency using the choose-max strategy based on the regional average energy, and the high frequency using the local-standard-deviation to being the basis of selective and weighted-averaging fusion method. This algorithm fully considered the image details and clarity, and fusion result is better. In addition, studies the selection of optimal decomposition level based on this method. Takes experiments on multi-focus images and medical images, and evaluate the result by objective and subjective evaluation, and obtains the optimal decomposition level.