SIA OpenIR  > 光电信息技术研究室
Alternative TitleResearch on Application Technologies Used in Image Coding Based on Compressed Sensing
Thesis Advisor赵怀慈
Keyword图像编码 压缩感知 抗干扰 图像重构 感兴趣区域编码 嵌入式编码 测量矩阵
Call NumberTP391.4/D78/2014
Degree Discipline模式识别与智能系统
Degree Name博士
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract实现复杂电磁环境条件下侦查监视图像高效、可靠、稳定和实时的编解码,是无线图像传输领域中的关键技术之一,是当前的研究热点。复杂电磁环境条件下的干扰主要分为随机干扰和突发干扰。本文主要针对自然干扰环境特性开展研究,旨在解决自然干扰环境条件下图像传输过程造成的兴趣区域纹理模糊、图像黑屏、马赛克等问题。相较于传统的图像压缩编解码方法,本文以自然干扰下的无线图像传输应用为背景,综合运用压缩感知、嵌入式图像编码等理论,研究抗干扰的图像压缩编码新方法,重点提高兴趣区图像编码和传输的实时性、增强图像编码算法抑制干扰能力,从而有效提高图像压缩编码性能。本文的研究成果具有一定的理论创新性和较高的应用价值。研究内容主要包括以下三个方面:1. 针对一般压缩感知测量矩阵的随机性导致生成复杂、计算复杂度高、占用存储空间过大等问题,在分析了确定性测量矩阵的特点和构造方法后,引入稀疏间隔构造概念,构造了一类基于Toeplitz矩阵的稀疏间隔测量矩阵。基于Toeplitz矩阵的稀疏间隔测量矩阵满足压缩感知理论的充分条件,能够实现图像的精确重构。实验结果验证了所构造测量矩阵的有效性。2. 为保证兴趣区域图像的优先编码和传输,基于分块压缩感知理论,借鉴嵌入式编码方法,提出基于压缩感知的位平面提升感兴趣区域编码和基于压缩感知测量值的感兴趣区域编码两种技术方案。方案一将位平面提升技术引入压缩感知,将压缩感知信号进行位平面分解后,对感兴趣区域位平面进行提升。方案二基于少量压缩感知测量值能够鲁棒重构图像的理论分析结果,将感兴趣区域压缩感知测量值按照需求进行提前编码。上述方案实现了图像兴趣区优先编码和传输,实验结果验证了方案的可行性和有效性。3. 针对自然环境干扰造成图像质量低的问题,提出一种基于压缩感知的图像压缩抗干扰重构技术,其基本思路为:首先通过检错码编解码确定受干扰压缩感知测量值的位置,然后剔除受干扰的压缩感知测量值,再依据剔除信息重新组织重构矩阵,最后进行图像重构。实验结果验证了所提出的抗干扰重构技术的有效性。
Other AbstractIt is one of key technologies in wireless image transmission area and a hot research topic that realizing efficient, reliable, stable and real-time coding and decoding to the images used for surveillance and reconnaissance under complex electromagnetic signal environment. The interference under the complex electromagnetic signal environment includes random disturbance and burst distrubance. This study mainly develops the research work aiming at the random disturbance in order to solve the problems that the decoded images have mosaic or blank screen. Aiming at the above problems, in order to improve the quality of the region of interest(ROI) and enhancing the anti-intefering ability of the image compression algorithm, and eventually to improve the performance of image compression, this dissertation takes the image transmission application under the wireless channel environment as the background, and comprehensively applies compressed sensing and embedded coding theories to the research of a new anti-intefering image ROI coding method. The elaborated research result provides an effective solution and has definite theoretical innovation and higher Application value.The dissertation develops the study work on image compression key technologies based on wireless channel from the construction of the measurement matrix, ROI coding method, and anti-interferring reconstruction technology respectively. The contents mainly consist of three aspects as bellow.Firstly, applying a measurement matrix to practice which only meets the theoretical demand has difficulty for the reasons of the randomicity of matrix elements, computation cost and storage cost. In order to solve this problem, a kind of measurement matrices is constructed based on Toeplitz. Because the theory condition the matrix should satisfy is a necessary but not a sufficent condition guiding the construction procedure, we should search for a kind of matrices by practice and analysis both satisfying theory condition and being easy to use. After analyzing the characteristic and construction means of the deterministic matrices, the notion of sparse interval is introduced, and a type of sparse matrices is constructed consequently. Experimental results proved that the constructed matrices are efficient.Secondly, on the problem that the bandwidth of the wireless channel is finite, in order to ensure the timeliness on ROI transmission, two ROI coding schemes are proposed based on blocked compressed sensing (CS) led by embedded coding theory. One scheme introduces the bitplane technology to CS, and shift up ROI bitplanes after decomposing all CS signals to bitplanes. The other scheme encodes the CS measurement values of ROI in advance based on the theoretical analysis that CS measurement values can reconstruct the image robustly. By above two schemes the intention that the ROI be coded and transmitted first can be achieved, consequently the compromise between the bandwidth restrict and the transmission real-time performance can be achieved, too. The experimental results indicate that both schemes are feasible and effective.Thirdly, if the image compressed by traditional image compression methods is disturbed when transmitted across the wireless channel, the loss of significant coefficients will cause the phenomenon that a part of the content of the reconstructed image is lost completely. For the reason that CS has robust reconstruction performance, an anti-interfering reconstruction technology based on CS is brought forward. The processing procedure first fixes the positions of the disturbed measurement values by means of error detecting codec, then eliminates these measurement values and reorganizes the relevant reconstruction matrix, finally reconstructs the image. On the basis of the second ROI coding scheme, the anti-interfering reconstruction experiment is conducted. The results of the experiment verify that the proposed method is valid.
Contribution Rank1
Document Type学位论文
Recommended Citation
GB/T 7714
杜梅. 压缩感知理论在图像编码中的应用技术研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2014.
Files in This Item:
File Name/Size DocType Version Access License
压缩感知理论在图像编码中的应用技术研究.(8854KB) 开放获取CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[杜梅]'s Articles
Baidu academic
Similar articles in Baidu academic
[杜梅]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[杜梅]'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.