【作者】 邹彦艳；

【导师】 曹广华；

【作者基本信息】 大庆石油学院 ， 地球探测与信息技术， 2004， 硕士

【摘要】 随着计算机和网络技术的普及，人们对数字图像在质量、大小和应用等方面提出了更高的要求，希望能够用有限的空间和带宽资源存储和传递大幅图像，并且根据实际需要，得到不同分辨率或质量的重构图像。这就要求图像压缩技术不仅有良好的压缩效率，而且还要能灵活地处理压缩码率。基于以上原因，对压缩编码方法的探讨和研究在近几十年来受到了广泛的关注。近20年来，在预测编码、变换编码和其它传统编码方法的基础上，图像压缩编码的理论、方法和技术取得了较大进展；其中小波编码算法因其能量集中特性、多分辨分析概念的提出和小波分解与重构快速算法的实现，使得小波变换在信号处理领域得到了广泛的应用，小波分析应用于图像数据压缩是小波分析应用的一个比较成熟的方面。 本文首先介绍了几种传统的压缩编码方法，并对国际标准JPEG和JPEG2000做了基本阐述，重点讨论了小波的多分辨分析、小波分解、重构算法、小波嵌入零树编码算法和改进的小波编码算法SPIHT。在此基础上提出了对图像进行预处理的想法，结合人眼对图像低频部分比对高频部分敏感的视觉特性对图像各频段提出不同的编码算法。本文在对图像的压缩过程中，对图像低频部分采用无损压缩方法(哈夫曼编码)，对高频部分采用改进的SPIHT编码方法，从而保证了在得到高压缩比的前提下使信噪比恶化程度较轻。本文对几种编码方法得到的重建图像进行了质量比较。最后，讨论了运动图像压缩核心技术——运动补偿技术，并用二维静止图像小波变换与一维时间轴上的小波变换相叠加的方法实现了对运动图像的压缩。更多还原

【Abstract】 With the popularization of computers and network technology, people brought forward higher requirements on the quality, size and application of digital images, hoping to store and transfer big images by using limited space and bandwidth, and to obtain images of various resolution ration or quality. This requires the image-compression technology to be not only more efficient in compressing images, but also to be more flexible in dealing with compress coding ratio. As a result of all the above, great attention has been paid to the discussion and research of code-compressing method in recent decades. During the latest 20 years, great progress has been made in the theory, method and technology of image-compression coding on the basis of such traditional coding methods as coding forecast and exchanger coding etc; wavelet transformation has been widely used in signal-disposal field, because it has a characteristic of energy focus, it puts forwards a concept of multi-resolution ration analysis and the realization of algorithm about wavelet decompose and fast rebuild. Among the applications of wavelet analysis, it is rather mature to apply it to image-data compressing. In the first place, this article introduces several traditional code-compressing methods, and makes a brief introduction about the international standard JPEG and JPEG2000, and dwells on the wavelet multi-resolutions dissecting, decompression and rebuild, EZW algorithm, improved SPIHT. Then, the idea to pre-treat images is put forward on the basis of what has just been discussed, combining with the visual characteristic that people are more sensitive to low-frequency than high-frequency parts of images, different coding algorithms are suggested to use for various frequency. During the image-compressing, lossless-compression method (Huffman-Coding) is used to deal with the low-frequency part, while SPIHT-coding is used to deal with the high-frequency part, so mat high compression scale can be achieved without worsening the SNR. In the article, the quality of the rebuilt images obtained by using different coding methods is compared. At last, the core technology of compressing moving images is discussed-motive reward technique, and the compression of moving images is realized by using the method that the two wavelet transformations of two-dimensional still images and one-dimensional time-axis are combined together.更多还原