Filtering Algorithm Research For Some Types Digital Images

In the process of acquisition and transmission,digital image tends to be disturbed byvarious noises,the noise,in the form of useless information,reduces the image quality andimpacts the successive work,such as image restoration,division,feature extraction andrecognition,and the treatment of noise is very important job in image processing．In recent years,many denoising algorithm appeared in the image denoising fields,andmade great achievements．For example,In the research of removing Gaussian noise,In 2005,person such as Morel J．M etc proposed the Non—local Mean Filter algorithm；In the researchof removing Impulse noise,In 2006,person such as Gamett etc proposed the Trilateral Filteralgorithm,In 2007,person such as Yiqiu Dong etc proposed the improved Trilateral Filteralgorithm,effectively improve the algorithm'S ability of detection of impulse noise,In 2009,such person as Li Bing who proposed a new method of removing mixture noise,which is alsoa good improvement of the Trilateral Filter algorithm．On the basis of previous studies,this thesis mainly includes two aspects:1,Based on thelinear gradient algorithm hybrid noise filter algorithm,Consider applying both gray valuedifferential and Gradient modulus value into judgrnent the correlation between the pixel size,we presents a new mixed noise linear denoising algorithm,which is shortened as W-LMF．This algorithm shows better denoising ability for low—level Gaussian noise image andcontains a low level of Gaussian noise'S mixed noise image,and most of the algorithm'Soptional parameter values can take for setting value,therefore the algorithm'S validity isimproved．2,In order to reduce the running time of the Non—local Mean Filter algorithm,Combining mean,variance and gradient these can reflect the similarity between pixels in thispaper,We presents a new fast algorithm．Nearly kept the original denoising algorithm'S results．Algorithm'S optional parameter values can take in a small range,thus reduced the parameterselection'S difficulty．The main content of this article is divided into six chapters:chapter one mainlyintroduces image and image processing,Digital image denoising technology'S Research statusat home and abroad,this article'S research significance and the text of the structurearrangement．chapter two mainly introduces digital image denoising theoretical basis,chapterthree mainly introduces Several existing related denoising algorithm,chapter four mainlyintroduces a new mix noise linear filter algorithm,chapter five mainly introduces a new fastalgorithm based on the non—local mean filter,chapter is the end of this article,summarizedthis paper'S job．...