The Study On Algorithm Of Lossless Image Compression Based On Wavelet Transform

The 21st century,an era full of information everywhere,an era of information explosion.Today,the Internet has become indispensable in people's lives,every year the amount of data information with rapid expansion,the rapid growth of the development of storage media has not complied with the increase of massive data,limited network bandwidth can't meet the people's demand for data transmission efficiency,which makes data compression naturally become an important research topic today.Image data is an important component of the information data member,the storage and transmission of image information in our lives more and more frequent,research compression coding technology for image data will gradually become the current research focus.Image compression is to remove the redundancy of image data,with fewer date to show the image same,thereby realizing more efficient data storage and transmission to data.Image redundancy:psychological visual redundancy,coding redundancy and a redundancy between the pixels.Remove the psychological visual redundancy by ignore the secondary information which visual unable to identify,so we can realize lossy compression by using this principle.The elimination of redundancy between pixels is through the image transform to remove redundant data between pixels.Coding redundancy is through appropriate coding techniques to achieve the purpose of as small as possible representing the image.The thinking of image lossless compression is through image transformation to remove the redundancy between pixels,then use the appropriate coding technique to encode the transformed image.As a kind of image processing tools,wavelet transform can well remove the redundancy between the pixels.The traditional wavelet transform decomposes the image into the pixel value which is irrational number,it is not suitable for lossless compression.While the integer lifting wavelet transform decomposes the image into the pixel value which is integer number,in lossless compression technology,it is the best choice for eliminate the pixel correlation.Based on the(5,3)wavelet as example analyzes the two-dimensional integer wavelet decomposition and recovery algorithm,(5,3)wavelet to decompose the image into a low frequency subimage and three high frequency subimages.The low frequency images reflect overview of the images,and the high frequency subimages reflect the image details.High-frequency pixel is by eliminating the correlation between the adjacent pixels.There are eight adjacent pixels around each pixel,while Two-dimensional integer lifting wavelet transform is only able to eliminate the correlation of three adjacent pixels.At the same time,and the red-black wavelet is also not fully eliminate the correlation between the adjacent pixels.So this thesis presents a new method.The method divided the image into a lot of 3 3 pixel collection,The center pixel as a low-frequency pixel,Eight high-frequency pixels can be obtained by eliminating the correlation between the surrounding pixels and the center pixel.In this way,image can be decomposed into a low frequency subimage and eight high frequency subimages,and fully eliminate the correlation is obtained.Through the comparison and analysis,the fact,the algorithm complexity of the method is the lowest,was found.And because of the proportion of the high frequency pixels and low frequency pixels,the new method can obtain higher compression ratio.The appropriate coding algorithm can make less encoded data.Through wavelet transform image,less low-frequency pixels and a lot of high-frequency pixel can be obtained.At the same time,the low frequency pixels' gray value range is large,while most of the high frequency pixels' gray value is small.Huffman coding is chosen as the coding algorithm to code the images after wavelet transform.In order to evaluate the performance of the new algorithm,selecting the Haar wavelet,wavelet(5,3),the red-black wavelet and the algorithm in this paper to transform image,then using Huffman coding it.The experimental data show that the performance of the new method is better than the other.So the new method is a better image lossless compression algorithm.