The Research Of Image Fusion Based On Wavelet Packet And Bset Tree

The purpose of mage fusion is getting a more exact and more distinguishable image for a region by distilling and integrating the information of multi-source images. In recent years, the technology of image fusion has played an increasingly important role in modern aerospace, automation, remote control and medicine, particularly in the field of military guidance. Image fusion includes 3 levels:pixels-level fusion,features-level fusion and decisions-level fusion from the abstract level.This paper is major on pixels-level fusion and emphasis on wavelet transform. At first, this thesis introduces the background, the development status home and broad of image fusion, and its problems at present.And then presents the conception of remote sensing image fusion. This paper summarized the evaluation methods of the RS image fusion, the evaluation methods include the subjective evaluation and the objective evaluation. Wavelet transform in image fusion has acquired better effects, wavelet analysis mufti-resolution analysis, Mallat theories algorithm, wavelet packet transform and best tree analyze are systematically reviewed in this paper. Base on confected images, we decomposes images using wavelet packet algorithm, This algorithm decomposes the images respectively, and makes a elaborate analysis on every image layer based on every previous image layers, and then processes subimages with matching methods, at last, we can get the fusion image by converse transform.Based on the wavelet packet algorithms, a new method is introduced which is based on the best tree transform in this paper. At first, the images are decomposed using the wavelet packet algorithm. Secondly, every subimage is decomposed based on the entropy theory, if the entropy of the sum of subcode is less then fathercode, the image can be showed by reconstructed and reverse transform. The value of entropy can scale the information of an image, so the best tree algorithm is wonderful according to the objective evaluation, and it also considers the standard of subjective evaluation. It is proved that it is a good methods than some classical fusion algorithm.