压缩采样技术和其应用.pdfVIP

第32 卷第2 期 电 子 与 信 息 学 报 Vol.32No.2 2010 年2 月 Journal of Electronics Information Technology Feb. 2010 压缩采样技术及其应用 金 坚 谷源涛 梅顺良 (清华大学电子工程系 北京 100084) 摘 要：如何降低宽带模拟信号数字化过程中的采样率，以及如何有效的对大量数据进行压缩存储一直是学者们关 心的问题。该文综述了最近出现的一种新型信号处理方法—压缩采样(Compressive Sampling, CS)，也称压缩传感 (Compressive Sensing)。该方法通过对稀疏信号进行观测而非采样，只需少量观测点就能精确的重构原始信号。结 果表明新方法的观测频率可以远远低于奈奎斯特采样频率。该文除介绍其基本原理和主要实现方法外，同时列举了 多种应用，并指出若干待研究的问题。 关键词：压缩采样；稀疏性；观测矩阵；信号恢复 中图分类号：TN919.5 文献标识码：A 文章编号：1009-5896(2010)02-0470-06 DOI: 10.3724/SP.J.1146.2009.00497 An Introduction to Compressive Sampling and Its Applications Jin Jian Gu Yuan-tao Mei Shun-liang (Department of Electronic Engineering, Tsinghua University, Beijing 100084, China) Abstract: The problems of how to reduce the sampling rate in the broadband analog signal digitization and how to compress effectively the large amount of data for storage are always concerned by researchers. The recent proposed Compressive Sampling or Compressive Sensing method to solve the said problems is introduced in this paper. The method, which employs non-adaptive linear projections that preserve the structure of the signal, can capture and represent the compressible signal at a rate significantly below Nyquist rate. This paper not only presents the key procedures of this theory but also lists a variety of applications and points out the questions to be studied. Key words: Compressive Sampling (CS); Sparsity; Measurement matrix; Signal reconstruction 带模拟信号采样率和提高海量数据压缩存储效率的 1 引言 新途径一直是学者们关注的问题。 香农/奈奎斯特采样定理指出：为避免信息丢 失，实现无失真恢复原始信号，采样率至少要两倍 对于稀疏信号，压