Comparison and Improvement of Wavelet Based Image Fusion

A multi-resolution fusion algorithm, which combines aspects of region and pixel-based fusion. We use multi resolution decompositions to represent the input images at different scales, and introduce multi resolution or multi modal segmentation to partition the image domain at these scales. This segmentation is then used to guide the subsequent fusion process using wavelets. A region-based multi resolution approach allows us to consider low-level as well as intermediatelevel structures, and to impose data-dependent consistency constraints based on spatial, interand intra-scale dependencies. The wavelets used in image fusion can be classified into three categories Orthogonal, Bi-orthogonal and nonorthogonal. Although these wavelets share some common properties, each wavelet also has a unique image decompression and reconstruction characteristics that lead to different fusion results. In this project the above three classes are being compared for their fusion results. Normally, when a wavelet transformation alone is applied the results are not so good. However if a wavelet transform and a traditional transform such as IHS transform or PCA transform are integrated for better fusion results may be achieved. Hence we introduce a new novel approach to improve the fusion method by integrating with IHS or PCA transforms. The fusion results are compared graphically, visually and statistically and show that wavelet integrated methods can improve the fusion result, reduce the ringing or aliasing effects to some extent and make image smoother.