基于MATLAB的JPEG基本系统编码_本科毕业论文.doc

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，经常采用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用Z形扫描，得到直流系数和交流系数。接着，对直流系数采用预测编码，对交流系数采用可变长编码。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、直流系数和交流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，采用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 ABSTRACT The 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 results of simulation demonstrate that there is no difference between reconstructed image and original image and reconstructed image can satisfy human visual requirements. Additionally, compression ratio is about 10 and peak signal-to-noise ratio is all over than 30dB. Therefore, the realization of JPEG basic system coding using MATLAB is of such advantages as simple method, fast speed, small error and it can enhance the efficiency and accuracy of image compression greatly. Key words: JPEG; discrete cosine transform; MATLAB; graphical user interface