Fractal-Based Image Processing Technology Of Amplification

The use of image interpolation can be achieved in different proportions for digital image magnification. Conventional amplification technique is based on Ngquist sampling theorem; both there are some defects which reduces the quality of the image in the image processing. As the infinite details of the fractal and its self-similarity, this paper adopts the method of fractal interpolation to enlarge the image processing to keep the image enlarged visual effects.First from digital image of the basic theory ,and from sampling points and quantitative series of view, the article elaborates on relations between the sampling points, the quantification series and the picture quality which if sampling points is more, the quantification series is higher, then the obtained picture quality is better. Also the article has analyzed the objective evaluation and the subjective evaluation criteria of the image quality, and has obtained that objective evaluation's characteristic is uses of the objective indicator and the quantitative index, but subjective evaluation's characteristic is the subjectivity and the qualitative evaluation. The most common assessment method is the subjective evaluation.Apart by introducing the basic theory and methods, the article advances IFS-based image magnification techniques which affine transformation to the sampling point to obtain the required image. Affine transformation need for the strict self-similarity and the fixed attractor so that it lacks randomness. Therefore that applicable to local enlarged image processing Local enlarged image processing.The fractal Brown function is the stochastic fractal mathematical model that can better reflect the statistic characteristics of image texture. This article uses the fractal Brown function in the image multiplication, and advances image interpolation method that is based on fractal Brown random field model——the random midpoint displacement method, and then realizes the programming with the MATLAB software.Finally for the different content image and the different picture element image, that uses the different enlargement technology to carry on enlargement processing separately in the MATLAB software .And Some images also carry on continual enlargement processing to confirm the technique validity.Through the analysis of experimental results, we can find images which draw random midpoint displacement method to achieve amplification of the image processing of images. Although there are some vague, but the details was a significant improvement and new images can maintain the original image texture.