Based On The Wavelet Transform Digital Image Mosaic Algorithm

Image mosaic technique is the one which can integrate image series from real world into a wide field of vision picture. It can eliminate the redundancies between image series; reduce memories of the information, making express information efficiently. Generally speaking, image mosaic is for realizing the smooth transition of the content of the image and eliminating the track owing to the variety of illumination. It needs to find a line between two images which will be dealt, and then makes the images into a smoothness image, which has high resolving power, either geometry or greyhound, geometrically and greyhound. The process includes geometric rectification and image registration which are essential, and then the cutting and mosaicing of image, color harmony will be the last step.There are some popular image mosaic methods such as statistic arithmetic, fuzzy set arithmetic and nerve network arithmetic. Orthogonal wavelet transform is also an efficient image merging method. Its special characteristics about localizing the time-frequency can distill information form signals and analyze the image by flexing and moving.This paper introduces the theory about wavelet and the applications on image processing particularly and gives two image merging ideas and gives two different boundary reconstruction methods.By using the special characteristics of the analogical orthogonal transform---biorthogna- l wavelet transform, a new method of image merging is presented. Under the conditions that high-pass filter and low-pass filter all have an odd number of non-zero samples and are symmetric about zero, there is a problem that the length of filter is bigger, the result of merging is better, but the boundary reconstruction is not satisfied. The problem is solved and the image is reconstructed precisely by using a new boundary extension method in this paper. The validity of the method is illustrated by the simulation results.