Research On Image Noise Inhibition Based On Fractional Order Lower Statistics

a stable distribution and fractional lower order statistics have been greatly concerned and extensive studied in recent years. It is widely used in the fields such as communication, radar, sonar, geological exploration, biomedical engineering and many other fields.Image denoising problem has always been a hot issue in the field of image processing. The traditional methods of denoising image are mostly designed basing on Gaussian noise. The facts prove that in the process of image processing and analyzing, the non-Gaussian features of some images, especially those speckle noise make the capability of traditional methods of signal processing degrade. Fractional lower order statistics (FLOS) is a key means of analyzing with the situation of non-Gaussian distribution image signal noise. This paper is mainly carried out on this topic, including the following:(1) Introduce the concepts and development status of a stable distribution and fractional lower order statistics, the significance of the study and the current applications in the filed of image processing.(2) Study and compare some traditional image denoising algorithms of their denoising results and their evaluating with the peak signal to noise ratio function under MATLAB simulation platform.(3) Research on image denoising algorithms based on fractional lower order statistics, this paper focuses on research and analyze adaptive filtering p-norm image denoising algorithms based on fractional lower order statistics.(4) Present a new adaptive filtering algorithm (A—RLMMN) based on several image enhancement algorithms, which is applied in a stable impulsive noise. According to the algorithms introduced and the new algorithm proposed, this paper carries on the simulation experiment under MATLAB simulation platform. Simulation results show the advantages and feasibility of the new algorithm.In a conclusion, the method based on fractional lower order statistics, has better toughness than the method based on Gaussian distribution. The method of adaptive filtering based on fractional lower order statistics, will indicate its superiority in image processing and it has in-depth research significance.