图像在网络上实现快速传输及实时处理技术.docVIP

图像在网络上实现快速传输和实时处理技术 1．JPEG压缩原理及特点 1. The JPEG compression principle and characteristics JPEG算法中首先对图像进行分块处理，一般分成互不重叠的 大小的块，再对每一块进行二维离散余弦变换（DCT）。变换后的系数基本不相关，且系数矩阵的能量集中在低频区，根据量化表进行量化，量化的结果保留了低频部分的系数，去掉了高频部分的系数。量化后的系数按zigzag扫描重新组织，然后进行哈夫曼编码。JPEG的特点如下： Firstly to block the image in the JPEG algorithm, generally divided into non-overlapping blocks, the size of the back of each block of two-dimensional discrete cosine transform (DCT). Basic not related coefficient after the transformation, and the coefficient matrix of the energy is concentrated in low frequency area, according to the quantization tables for quantitative and quantitative results retained the coefficients of low frequency part, get rid of the coefficient of the high frequency part. After quantitative coefficient according to reorganize the zigzag scan, and then to Huffman encoding. JPEG features are as follows: 优点：（1）形成了国际标准；（2）具有中端和高端比特率上的良好图像质量。 Advantages: (1) formed the international standard; (2) has the midrange and high-end bitrate of good image quality. 缺点：（1）由于对图像进行分块，在高压缩比时产生严重的方块效应；（2）系数进行量化，是有损压缩；（3）压缩比不高，小于50[2]。 Shortcomings: (1) because of image block, the high compression ratio have serious tetris effect; (2) quantify the coefficient of, is a lossy compression; (3) the compression ratio is not high, less than 50 [2]. JPEG压缩图像出现方块效应的原因是：一般情况下图像信号是高度非平稳的，很难用Gauss过程来刻画，并且图像中的一些突变结构例如边缘信息远比图像平稳性重要，用余弦基作图像信号的非线性逼近其结果不是最优的[3]。 JPEG compressed images appear square effect: is the cause of general image is a highly nonstationary signal, using Gauss process to depict very hard, and some mutations structures such as edges in the image information is far more important than the image stability, make image signal of the nonlinear approximation with cosine matrix [3] the result is not optimal. JPEG压缩 JPEG compression 负责开发静止图像压缩标准的“联合图片专家组”（Joint Photographic Expert Group,简称JPEG），于1989年1月形成了基于自适应DCT的JPEG技术规范的第一个草案，其后多次修改，至1991年形成ISO10918国际标准草案，并在一年后成为国际标准，简称JPEG标准。 Is responsible for the developmen