Dissertation > Excellent graduate degree dissertation topics show

Salt and pepper noise in digital image filtering algorithms

Author: GuoHongWei
Tutor: YuJiang
School: Yunnan University
Course: Communication and Information System
Keywords: Salt and pepper noise Median filter Histogram Matlab
CLC: TP391.41
Type: Master's thesis
Year: 2010
Downloads: 247
Quote: 2
Read: Download Dissertation

Abstract

The digital image is formed, transfer and recording process, due to the imperfections of the imaging system, the transmission medium and the recording apparatus tends to make the image acquired by a variety of noise pollution. Salt and pepper noise image quality damage, even a small amount of noise will be greatly damaged the image details. Therefore, in the field of pattern recognition, computer vision, image analysis and video coding, image pre-processing of the noise is extremely important, the treatment effect is good or bad will directly affect the quality and results of the follow-up work. Image denoising purpose is to diminish the noise and improve the signal-to-noise ratio, and to protect the edges and details. In this regard, the non-linear filter is more effective than the linear filter. Traditional median filter as a nonlinear filter based on statistical sort, can effectively eliminate the salt and pepper noise in the image, while protecting the edges and details. In addition, because the median filtering technology is easy, fast, so widely used. In order to improve the filtering performance, have been developed many modification of the median filtering algorithm. This paper first introduces the noise model filter the basics. Then, the analysis of the research progress of the salt and pepper noise filtering, summed up the current filtering technology switching technology, technology of weighted median and weighted mean, and some important the algorithm detailed description and experimental comparison. Finally, three new algorithms are based on extreme value and line detection salt and pepper noise filtering algorithm based on adaptive histogram information to remove salt and pepper noise algorithm, adaptive control iteration random value impulse noise filtering algorithm. Among them, the salt and pepper noise filtering algorithm based on extreme value and line detection filter out salt and pepper noise to protect the image pixel black, white thin line. Ideal model for the salt and pepper noise filtering to remove the salt and pepper noise algorithm can get better filtering effect than most algorithms based on adaptive histogram information, the filtering process does not need to set the parameters, strong adaptability to noise rate of 0.8 0.9 noise pollution image can be restored. Adaptive iterative random value impulse noise filtering algorithm switch filter ideological impulse noise characteristics of random values, the algorithm is divided into noise detection, noise filtering, false detection pixel correction and iterative control four steps, the new algorithm is more suitable for random A value of impulse noise filtering. Experiments, the filter out random value impulse noise, the algorithm filtered image detail better information protection, the filtering process does not need to set the iteration parameters, adaptive median filtering algorithm.

Related Dissertations

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net Mobile