摘 要本文介绍了基于 MATLAB 的 JPEG 基本系统编码。在图像传送过程中，常常采纳JPEG 格式对静止图像进行压缩编码。JPEG 基本系统编码首先把灰度图像分成 8×8 的像素块，然后对各个像素块进行离散余弦变换得到变换系数后再进行量化。其次，对量化后的变换系数采纳 Z 形扫描，得到直流系数和沟通系数。接着，对直流系数采纳预测编码，对沟通系数采纳可变长编码。最后，根据标准的 Huffman 编码进行熵编码，输出压缩图像的比特序列，从而实现图像压缩。在接收端，经过 Huffman 解码、直流系数和沟通系数可变长解码以及反量化后，再进行离散余弦逆变换后得到重建图像。MATLAB 仿真结果表明：重建图像与原始图像几乎没有任何差异，能够满足人们的视觉需求。另外，数据压缩比在 10 倍左右且峰值信噪比均在 30dB 以上。因此，采纳MATLAB 实现 JPEG 基本系统编码具有方法简单、速度快、误差小等优点，能够大大提高图像压缩的效率和精度。关键词：JPEG；离散余弦变换； MATLAB；图形用户界面ABSTRACTThe JPEG basic system coding based on MATLAB is introduced in this paper. The JPEG format is usually used to compress static image during the process of image transmission.The JPEG basic system coding divides the gray image into several sub-images of size 8×8 firstly. Discrete cosine transform is used to get the transform coefficient of sub-image and then the transform coefficient is quantized. Secondly, Z type scan is adopted to get direct current (DC) coefficient and alternate current (AC) coefficient of the quantized transform coefficient. Thirdly, predictive coding and variable-length coding is used for DC and AC coefficient respectively. Finally, bit sequences of the compressed image are outputted by using entropy coding according to standard Huffman coding. Then image compression is realized. The user gets the reconstructed image by Huffman decoding, variable-length decoding of DC and AC coefficient, dequantization and reverse discrete cosine transform sequentially.MATLAB resul...
