Image Fusion Based On Wavelet Transform

Multi-sensor image information processing system is a tendency of the development of information technology and also is the result of emergence of many different advanced image sensors. Image fusion is one of the most important branch of multi-sensor data fusion. It defines the process of fusing visual information obtained from multi-sensors surveying the same region into a new, single fused image. The fused image has some specialities which cannot be obtained from a single sensor. The purpose of image fusion processing is to fully use the complementary and redundancy information provided by different image sensors in order to improve the performance of information processing system and to implement the detection, features abstraction and recognition of a target with better methods. Among fusion methods of different kinds, the Pixel-level Fusion method, which can best keep information from original images, is the prevalent method among professionals and the major study of this dissertationThe major contents of this dissertation are multi-sensor image fusion technology based on wavelet transform and use it in medical image fusion and multi-focus image fusion. First we introduce the concepts, advantages, developments and applications. Three levels of image fusion and image fusion techniques in common use are also reviewed. In order to estimate the fusion result, summarizing the subjective and objective evaluation criterion for the performance of image fusion.The wavelet transformation elementary theory and the general wavelet fusion principle are introduced too. The dissertation points out the disadvantage of the traditional algorithms that the importance of the fusion algorithms. This dissertation utilize biorthogonal wavelet transform merge medical image for avoiding phase distortion by conventional orthogonal wavelet transform, and introduce a new fusion method. Biorthogonal wavelet gives up the orthogonal character to acquire the symmetry and compactness. Compare to orthogonal wavelet, biorthogonal wavelet figure have more choice scope, so it is more flexible to be design. This dissertation researches the system functions correspond to focused image and defocused image based on imaging mechanism of panchromatic sensor and propose a novel effectively multi-focus image fusion method. Because orthogonal wavelet transform is short of translation invariant and have vision distortion. Wavelet redundancy algorithm is used to decompose the original images in different frequency region. Redundancy algorithm is to interpose proper zeros between arbitrary points of filter beforethe filters h₀(n) and h₁(n) convolve with low frequency image. While choosing highfrequency coefficients, we use spatial correlation filter method which could inhibit noise and select the pixel which has strong correlation and prominence character as the fusion image wavelet coefficient.