高噪声背景下舰船指挥舱大词汇量连续语音识别方法

Large vocabulary continuous speech recognition method for ship command cabin in high noise background

传统的舰船指挥舱大词汇量连续语音识别方法存在着识别错误率高的缺陷,为此提出高噪声背景下舰船指挥舱大词汇量连续语音识别方法研究。对采集的连续语音信号进行预加重和预处理,以预处理后的连续语音信号为基础,采用多通道语音增强方法对连续语音信号进行增强,得到纯净连续语音信号估计,采用CDMFCC方法对纯净连续语音信号特征参数进行提取,通过CDHMM方法实现了高噪声背景下舰船指挥舱大词汇量连续语音的识别。通过实验得到,提出的舰船指挥舱大词汇量连续语音识别方法识别错误率比传统方法低了16%,说明提出的舰船指挥舱大词汇量连续语音识别方法识别性能更好。

The traditional method of large vocabulary continuous speech recognition for naval command cabin has the defect of high recognition error rate.Therefore,a method of large vocabulary continuous speech recognition for naval command cabin in high noise background is proposed.The acquired continuous speech signal is pre-emphasized and pre-processed.Based on the pre-processed continuous speech signal,the multi-channel speech enhancement method is used to enhance the continuous speech signal,and the pure continuous speech signal is estimated.The feature parameters of the pure continuous speech signal are extracted by CDMFCC method,and the command cabin of naval ships under high noise background is realized by CDHMM method.Large vocabulary continuous speech recognition.The experimental results show that the recognition error rate of the proposed method is 16%lower than that of the traditional method,which shows that the proposed method has better recognition performance.