Research Of The Rule On Wavelet Coefficient-based Image Fusion

Image Fusion technique has the effect that can transact the image that multi-channels collects, these images come from the same object. The image fusion can extract each of channel information and integrate into one image or to be observed for further processing. Image fusion can integrate multi-images from the same object into one picture, and fused image can be easily understood than each primary image and be close to reality object. Image fusion has a broad application in sensor, chronometer, physics, medicine, bioscience and info-science.Wavelet analyze has very important application in image processing, including image compression, image remove-voice, image retrieval, image fusion and video stream. Image fusion is a very active research area in domestic and abroad in recent year, and takes very important consequence in image processing. The application of 2-Dimension wavelet analyze is an important aspect, and the image fusion based on wavelet analyze turns into a successful application area of wavelet theory.At first, the dissertation introduces the background, the development conditions of image fusion and widely applications. Then the theory of wavelet analysis is introduced in detail, especially in DWT and multi-wavelet decomposition,thesse theory and technique are foundation of image fusion. We also introduce kinds of image fusion technique and probe into the evaluation standard of fusion effect. The paper introduce by the number application of 2-D wavelet analyze in image fusion and research the method based wavelet of fusing different image deeply and widely, including fusion frame based wavelet transform,the advantage of wavelet decomposition comparing HIS transform and effect influence of multi-wavelet decomposition aiming at image fusion. The dissertation brings forward a new kind of arithmetic based on wavelet image fusion. Mass of experiment data validates the validity and the feasibility of the algorithm that the dissertation proposes. Aim at multi-focus image and multi-sensor image, the imitation experiments prove that our algorithm offers a good performance in obtaining the background character information of different original images .Original image detail is well retained.