JPEG2000图像压缩关键技术研究及FPGA实现

	JPEG2000图像压缩关键技术研究及FPGA实现
其他题名	Study on the Key Technologies of JPEG2000 Image Compressionand FPGA Implementation
	陈占良
学位类型	硕士
导师	金龙旭
	2015-10
学位授予单位	中国科学院大学
学位专业	光学工程
关键词	图像压缩 Jpeg2000 小波变换 Tier2 编码 Ddr3 存储器
摘要	遥感卫星是目前空间对地观测的最有效方法，也是衡量一个国家航天技术的重要指标。随着我国基础工艺的发展，遥感图像的分辨率越来越高，从而使存储和传输图像的数据量越来越大。由于遥感卫星有效载荷和传输带宽的局限性，有必要对图像数据进行压缩处理，减少图像存储所需介质和传输时间。在图像压缩技术中，有三种主要的数据冗余可以进行压缩处理：像素间冗余、编码冗余和心理视觉冗余；减少或者去除这三种冗余中的一种或者几种都可以实现图像的压缩。小波变换具有多分辨率的特性，能够根据人眼的不同视觉特性提供不同分辨率下图像的细节信息，而且小波变换后的信息包括空域和频率两方面，对图像压缩提供了便利。采用基于小波变换的JPEG2000 标准协议对遥感图像进行压缩处理，首先对图像数据进行小波变换，去除图像像素间的相关性，然后对变换后的小波系数量化处理，采用最优截断的嵌入式块编码算法对小波系数进行熵编码，减少图像数据的编码冗余，最后通过tier2 编码器输出符合压缩协议的标准码流。JPEG2000 图像标准协议具有高压缩比、高信噪比的特点，能够同时实现有损压缩和无损压缩，由于采用的是嵌入式码流的编码式，因此可以实现图像的渐进传输。虽然有这么多的优点，但是由于JPEG2000 标准协议的编码算法复杂，并且需要大量的存储资源。因此，本文提出了一种软硬件相结合的方式来实现JPEG2000 编码算法中的tier2 编码器，首先使用verilog HDL 构造一个硬件CPU，再用软件实现tier2 编码算法，最终将软件编码嵌入到硬件CPU 中运行实现tier2 编码。在保证运算速度的要求下，很大程度上减小了硬件实现的复杂度。由于小波变换要求的图像数据是无限长的，但是实际的遥感图像都是有限的，因此在做小波变换时就必须对图像的边界进行处理。JPEG2000 标准协议采取的方法是对边界数据进行周期对称延拓，但是对于要求实时压缩的遥感图像来说，周期对称延拓就会引入时间延迟，从而降低整个系统的工作效率。因此，本文抛弃边界数据的周期对称延拓的方法，对边界数据单独进行运算，严格控制边界数据流，从而减少了数据运行时间。压缩过程中会产生大量的中间数据，采用双倍数据速率随机存储器DDR3 进行数据的缓存传输，可以满足大数据量的高速存储传输。
其他摘要	Remote sensing satellite is the most effective way to observe the earth from the space, and it is a important norm to judge the national space technology. With the development of basal technology, the resolution of remote sensing image is more andmore high, which results in a lot of data needing to be stored and transferred. But, the storage medium and bandwidth are limited. So, it is essential to compress the remote sensing image. There are three kinds of data redundancy in image, including coding redundancy, pixel redundancy and psychological vision redundancy. If we can reduce or dispel one or more redundancy, we can compress the remote sensing image. There are loss compression and lossless compression, due to the different algorithm. Lossless compression can recover the image absolutely, without any distortion. The lossless compression can increase the compression ratio, but with the big distortion. Wavelet transform has multiresolution, and can provide different image information under the different resolution. The paper proposes the JPEG2000 which base on wavelet transform to compress the remote sensing image. Firstly, wavelet transform dispel the dependency among pixels. Secondly, to quantize the coefficient of wavelet transform. At last, to entropy coding the quantized wavelet transform with embedded block coding with optimized truncation, which can reduce the coding redundancy and output the standardized bit stream. JPEG2000 is characterized by high compression ratio and high signal to noise ratio, and it can realize loss compression and lossless compression at the same time. And it adopts the embedded bit stream coding, so it can transform the image data gradually. The algorithm of JPEG2000 is complicated, and it needs a lot of storage medium. The paper proposes a method to realize the entropy coding, which includes hardware and software. Firstly, it is necessary to structure the CPU using the verilog HDL. Secondly, using the C language codes the entropy coding. At last, to debug the CPU with embedded C language. On the premise of arithmetic speed, it reduces the complexity of the hardware design. The image data of wavelet transform should be limitless. In fact, all the image data is limited. So, it is essential to take measures to deal with the boundary image data. The paper abandons the symmetrical prolongation of boundary image data, and operates the boundary image data independently, which reduces the operation time. During the image compression, there are lots of transitional data need to be stored. The paper uses DDR3 memory to store the data, which can satisfy the needs of high speed transmission.
语种	中文
文献类型	学位论文
条目标识符	http://ir.ciomp.ac.cn/handle/181722/49274
专题	中科院长春光机所知识产出
推荐引用方式 GB/T 7714	陈占良. JPEG2000图像压缩关键技术研究及FPGA实现[D]. 中国科学院大学,2015.

条目包含的文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
陈占良.pdf（5154KB）	学位论文		开放获取	CC BY-NC-SA	请求全文