语音信息查询虚拟仪器的设计

在虚拟仪器开发软件LABVIEW平台上，利用LABVIEW和MATLAB混合编程，通过声卡获得原始语音信号，经过小波消噪处理、预加重处理和端点检测处理得到干净的语音信号，提取语音信号的美尔频率倒谱系数及其一阶、二阶差分系数作为语音识别的特征参数，通过矢量量化（VQ）与隐马尔可夫模型（HMM）实现非特定人连续语音的训练与识别，构建了基于LabVIEW平台的连续语音识别系统．实验表明，系统的识别速度较快，识别率达到90％左右，而且成本比较低廉，具有一定的应用价值．

The Implementation of Continuous Speech Recognition System Based on LabVIEW

Based on LABVIEW and obtaining the original sound signal through sound-card,and by obtaining the clean sound signal through Pre-emphasis process,wavelet eliminating noise process and port detection process,a continuous speech recognition system is realized through VQ(Vector Quantization)and HMM(Hidden Markov Model) for training and recognition.The Mel frequency cepstrum coefficient and its one step,two step difference are used as speech recognition's characteristic parameter,and the continuous speech recogn...