IOE OpenIR  > 光电技术研究所博硕士论文
大面阵CCD图像实时压缩存取方法的研究与实现
Alternative TitleStudy and Implementation on Real-time Compression and Access Methods of High Resolution CCD image
陈苏婷
Subtype博士
Thesis Advisor吴钦章
2007-05-31
Degree Grantor中国科学院光电技术研究所
Place of Conferral光电技术研究所
Degree Discipline信号与信息处理
Keyword图像压缩 整数小波变换 嵌入式零数编码 硬件实时实现 快速存取
Abstract图像压缩作为信息社会必不可少的基本技术,经过几十年的发展,取得了巨大的进展,且在许多领域得到了成功的应用。本文通过对图像压缩算法进行深入研究,针对航测领域的特殊需求,提出了三种压缩编码方案。通过对这几种三种方案的比较分析,选取了其中一种作为大面阵CCD图像的实时压缩方案,并对该方案的硬件设计和实现进行了探讨,此外还针对大面阵CCD图像数据的快速存取和浏览方法进行了研究。 本文介绍了图像压缩模型以及评价图像压缩算法的原则,分析了常用的图像压缩标准,对无损压缩编码和有损压缩编码两种图像编码方式进行了比较,并概述了常见的无损压缩编码和有损压缩编码。 针对目前流行的整数小波变换编码方法作了深入研究,在对大量小波变换系数统计分析的基础上,提出了一种基于提升框架的子带图像混合编码方案,该方案针对变换后的小波系数在最低频子带和其余高频子带的不同特点,分别采取了不同的编码策略。仿真结果表明,该混合编码方法是一种高效的无损图像压缩编码方法,且具有很高的压缩比。 通过对EZW算法和SPIHT编码算法进行深入研究,提出了一种嵌入式零树结构的图像压缩编码方案。该方案通过对SPIHT算法的输出比特平面进行游程编码方法,很好的提高了图像压缩比和压缩效率。 结合SPIHT算法和LZC算法两者的优势,提出了一种基于LZC算法的图像实时压缩编码方案。该方案通过采用IB-IWT算法的经典代表(5,3)整系数小波变换,大大降低了计算复杂度和存储空间;同时,整合LZC算法的优点,对小波变换系数用三个标志位图代替原始SPIHT算法的三个大容量链表,大大降低了内存需求量;并将原始SPIHT算法中比特平面串行输出的结构方式改为比特平面并行结构输出,使重要性信息与幅值位信息同时处理,图像处理时间仅与分辨率有关,提高了编码速度和效率。 通过对上述三种压缩方案的详细分析,结合硬件特点,对基于LZC算法的图像实时压缩编码方案的硬件实现进行了探讨。从整个图像实时数据处理的硬件平台构架到整个算法各个部分的硬件设计与实现作了分析。提出了基于FPGA的(5,3)整系数小波变换快速实现方法,并针对小波变换的边界点处理、有限字长效应分析和小波系数压缩编码算法的硬件设计进行细致的探讨。最后,给出了整个压缩算法的硬件实现结构。 论文最后根据实际需求,对大面阵CCD图像的存取系统进行了研究,提出了大面阵CCD图像数据的快速存取和浏览方法。通过引入文件概念,把一次飞行下来的整个海量CCD图像数据组织为一个文件,并提出了相应的快速存取方法。并针对我室研制成功的高速图像数据记录系统,给出了该存取方法在系统中的实现结构。 本文针对航测需求,从理论上对图像数据压缩算法和快速存取方法进行了深入研究分析,提出了适合硬件实时实现的大面阵CCD图像压缩方案和快速存取方法,并给出了其实现结构。本文的工作作为对国内首台自主研制出的超高分辨率CCD相机图像实时压缩存取方法的研究与实现,具有很高的理论意义和工程实用价值。 关键词:图像压缩 整数小波变换 嵌入式零数编码 硬件实时实现 快速存取
Other AbstractThe implementation on real-time compression and access of high resolution CCD image is an urgent question in aerial photography and measurement. Although there are a lot of achievements in image compression algorithm, the general compression methods haven’t achieved the best compression field, therefore developing a real-time compression method which is suit for high resolution CCD image, especially for the first highest resolution CCD camera developed by our institute in country are of significant interest both in practice and theory. In this Ph.D diseertation, image compression module and the principle of image compression value is introduced; image compression standard is analyzed, pointing at two kinds of image compression coding methods (Lossless coding algorithm and Lossy coding algorithm), classical iamge compression algorithms are discussed. Through comprehensive research on integer wavelet transformation and vast statistics on wavelet transformation coefficients, sub-band hybrid coding method by lifting scheme is proposed. By using different coding methods in different sub-bands, compression ratio is increased by this method. The experimental result shows this method is high efficient lossless compression algorithm. By thorough study on embedded zero-tree wavelet coding algorithm,an embedded zero-tree coding algorithm is proposed. Through adopting run length coding for bit-plane output by embedded zero-tree algorithm, compression ration and efficiency are improved greatly. A novel compression algorithm based on LZC algorithm is proposed which takes advantages of both SPIHT algorithm and LZC algorithm, but achieves the better performance of SPIHT using the reduced memory requirement based on LZC. By using three flag maps instead of three mass ordered lists, memory requirement is reduced significantly; by operating significant map symbols and successive aprproximation symbols simultaneously, parallel output structure of bit plane is adopted compared to serial output structure of original SPIHT algorithm. From above improvements, coding speed and efficiency of the algorithm are increased and the algorithm is suitable for real-time hardware implementation in theory. From thorough analysis to the above three compression algorithm, from the whole image real-time processing platform structure to corresponding parts of hardware implementation about the compression algorithm based on LZC algorithm is discussed. (5, 3) lifting IWT (Integer Wavelet Transformation) based on FPGA is analyzed and realized. Meanwhile, boundary processing and finite word length effect is discussed in wavelet transformation. At last, the whole algorithm implementation structure is given. According to practical needs, data access system of CCD image is analyzed. From the point of document, intensive research on access method is carried out and a proper, fast access method is presented for high resolution CCD image data. Meanwhile, aiming at high speed image recording system developed by our teamt, the realization structure on the system is constructed.Through the realization of fast access method, the search efficiency and data processing speed are significantly enhanced. Obviously, they are of great engineering value. In this Ph.D dissertation, as application research on real-time data processing (real-time compression and fast access) about high resolution CCD image in aerial photography and measurement, not only image compression algorithm and fast data access method is researched in theory, but also the study on the technic of hardware implementation of real-time data processing algorithms is presented. Based on the intensive and thorough study both in theory and practice, the principle and realization of the whole system are designed. As the application research on the first super highest resolution CCD camera developed by our country, this Ph.D dissertation has great value both in theory and engineering. Keywords: image compression, integer wavelet transformation, embedded zerotree coding algorithm, hardware implementation, fast access.
Pages131
Language中文
Document Type学位论文
Identifierhttp://ir.ioe.ac.cn/handle/181551/223
Collection光电技术研究所博硕士论文
Recommended Citation
GB/T 7714
陈苏婷. 大面阵CCD图像实时压缩存取方法的研究与实现[D]. 光电技术研究所. 中国科学院光电技术研究所,2007.
Files in This Item:
File Name/Size DocType Version Access License
10001_20041801512701(3539KB) 开放获取CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[陈苏婷]'s Articles
Baidu academic
Similar articles in Baidu academic
[陈苏婷]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[陈苏婷]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.