Research On Image Filtering Method Based On Fuzzy Set Theory

Digital images are often corrupted by the various noise in the process of generation, acquisition and transmission, which will lead to image quality decline. This may result in difficulties in the subsequent image processing, such as edge detection, feature extraction,etc.Therefore, in the image processing , Image filtering is one of the very important to remove the impulse noise from the image. there are many types of Image noise the gauss noise and impulse noise are the two most commonly used noise models. This thesis is mainly concerned in the development on the filtering algorithms for the impulse noise based on fuzzy set theory.There are some image-denoising algorithms, such as median filter,mean filtering and its improved algorithms, which are normally considered to have good performances on removing impulse noise, but the median filtering algorithms normally have poor performances on protecting the image details. Recently, Fuzzy technique has been shown to be very well suitable to the noise of uncertainty, analyses its characteristics , many filtering algorithms based on fuzzy set theory are proposed.On the basis of the research of many image-denoising,this paper firstly introduces the correlation theory of fuzzy set theory and some classic image filtering algorithms, and then some well-established fuzzy filters are comparatively analyzed in this thesis. The experimental results demonstrate that he fuzzy filtering algorithms can remove impulse noise, and preserve more detailed information. The emphasis of this paper is as follows: Firstly, Impulse noise removal from digital images by a image-denoising algorithm based on type-2 fuzzy logic. Interval type-2 fls and gaussian type-2 fls are described in detail. Based on the combination of interval type-2 fuzzy systems and gaussian type-2 fuzzy systems, a novel filter is presented for filtering impulse noise in images. At last, Based on analyzing the principles of noise detector and noise filtering, an effective image denoising algorithm is proposed to restore images corrupted by impulse noise. The proposed algorithm utilizes the directional difference to decompose the window into four subwindows in order to distinguish accurately noise points from signal points by comparing the absolute weighted mean value of the differences between the center pixel and its neighboring pixels in four subwindows with an predefined threshold. And then according to the directional correlation-dependent, the proposed algorithm adopts a edge-preserving filtering method to reconstruct the value of the corrupted pixel. Finally, A lot of relative programs are designed by matlab to verify the properties of these algorithms,the experimental results show that the method proposed in this paper can obtain higher psnr value and preserve more detailed information.