毕业论文设计：盲信号处理.docVIP

学院毕业论文设计 盲信号处理 院 （系）：电子信息系 专业：电程技术 班级： 学号： 姓名： 指导教师： 目录 摘要 4 ABSTRACT 5 第一章 语音信号及噪声概述 1 1.1 语音信号的概述 1 1.1.1 语音特性分析 1 1.1.2 语音信号的基本特征 2 1.2语音噪声特性分析 3 1.2.1 信噪比(Signal Noise Ratio，SNR) 3 1.2.2 信干比(signal-to-Interference Ratio，SIR) 4 第二章 盲信号处理 5 2.1盲信号处理的概述 5 2.1.1 盲信号处理的基本概念 5 2.1.2 盲信号处理的方法和分类 5 2.1.3 盲信号处理技术的研究应用 6 2.2 盲源分离法 7 2.2.1 盲源分离技术 7 2.2.2 盲分离算法实现 7 2.2.3 盲源分离技术的研究发展和应用 8 2.3 独立成分分析 9 2.3.1 独立成分分析的定义 9 2.3.2 ICA的基本原理 10 2.3.3 本文对ICA的研究目的及实现 12 第三章 盲语音信号分离的实现及抑噪分析 15 3.1 盲语音信号分离的实现 15 3.1.1 盲信号分离的三种算法 15 3.1.2 不同算法的分离性能比较 17 3.2 抑制噪声的算法仿真及结果分析 17 3.2.1 抑噪算法仿真实现 17 3.2.2分离结果分析 20 3.2.2 不同算法的分离性能比较 28 第四章 结论与展望 35 致 谢 37 参考文献 38 附 录 39 盲信号处理 摘要 ABSTRACT Blind separation of speech signal processing means is the use of BSS techniques microphone detects a voice signal processing. Separation of mixed speech signals is important for blind separation, the current separation of overlapping speech is built mostly in the absence of aliasing noise in the environment, under the circumstances, mainly in blind source separation (Blind Source Separation, BSS), the statistical characteristics of signal Several observations from the unknown signal to recover the independent source components; This article focuses on the background to the blind speech signal processing method, signal processing, speech and hearing, how the noise from a mixture of overlapping speech signal to separate the various audio source signals to mimic human speech separation ability become an important research question. Combination of concrete realization of the main ICA technology, speech denoising as a preprocessing of speech with noise-aliasing blind separation were studied, the paper introduces three kinds of FastICA algorithm: SOBI algorithm and the CICA algorithm, three kinds of algorithm application the actual noise in speech signal separat