论文部分内容阅读
统计相关源信号分离理论不仅有着广泛的应用背景,也为深入了解数据的本质结构提供了有效的分析工具.首先,重点分析和讨论一类特殊的相关源信号分离模型——独立子空间分析模型的可分离性;其次,分别介绍基于源信号稀疏性、统计测度、独立子空间分析、源信号时序结构、源信号有界性和非负性的各种相关源信号分离算法;再次,通过将加性噪声中的盲源分离和高光谱解混问题建模为统计相关源信号分离模型,表明了该方法的应用价值;最后,总结了相关源信号分离中存在的问题,并对下一步的研究思路进行了分析和展望.
Statistical correlation source signal separation theory not only has a wide range of application background, but also provides an effective analysis tool for in-depth understanding of the essential structure of the data.Firstly, this paper focuses on the analysis and discussion of a special class of correlated source signal separation model-independent subspace analysis model Secondly, we introduce various source signal separation algorithms based on source signal sparsity, statistical measure, independent subspace analysis, source signal timing structure, source signal boundedness and nonnegative, respectively. Thirdly, The model of blind source separation and hyperspectral unmixing in additive noise is modeled as statistical source signal separation model, which shows the application value of this method. Finally, the problems existing in the separation of source signals are summarized, Research ideas were analyzed and prospects.