基于去噪技术的DSP语音识别系统设计

针对语音识别系统对抗环境噪声的实际需求,提出一种二次组合抗噪技术,研究并设计了一种以数字信号处理器(DSP)为硬件平台,以隐马尔可夫模型(HMM)为算法的抗噪声嵌入式语音识别系统.DSP采用型号为TMS320VC5509A的芯片,配以外围硬件电路构成语音识别系统的硬件平台.软件设计以离散隐马尔可夫模型(DHMM)为识别算法进行编程,系统软件主要有识别、训练、学习和USB四个主要模块.实验结果表明:基于二次组合去噪技术的语音识别系统有更好的抗噪声效果.

Design of DSP speech recognition system based on denoising technology

Aiming at practical demand for speech recognition anti-noise,propose a secondary combination antinoise technology,design an anti-noise embedded speech recognition system,which uses digital signal processor (DSP) as hardware platform and hidden Markov model (HMM)as algorithm.The hardware platform of speech recognition system is constructed by TMS320VC5509A DSP,and the system is matched with peripheral hardware circuit to form speech recognition system.Using discrete hidden Markov model(DHMM) as recognition algorithm for software design and programming,system software include four modules,which are identification module,training module,learning module and USB module.Experimental results show that speech recognition system based on secondary combination anti-noise technology has better anti-noise effect.