小波包在车载移动电话语音信号提取中的应用

Application of Wavelet Packet in Voice Signal Extraction of Mobile Telephone on Automobile

利用小波包理论对车载移动电话的接收信号进行小波包变换,通过设定一合适的阈值对变换后的信号进行量化处理,提取出主要由发动机产生的噪声信号,然后用实际检测信号减去小波包变换信号,得出汽车司机的语音信号,从而达到消除噪声的目的。利用MATLAB的小波工具箱对所提出的方法进行验证,结果表明提取后的信号与驾驶员的声音信号十分相似,误差较小,对发动机噪声有明显的抑制作用。

Wavelet packet transform is conducted in the reception signal of mobile telephone on automobile by using wavelet packet theory. The transformed signal is quantified and processed by setting up a proper threshold, the noise signal generated by the engine is extracted, and by subtracting the wavelet packet transform signal from the detected practical signal, the speech signal of the auto-driver is obtained, and thereby the noise is eliminated. This method is verified by the wavelet toolbox of MATLAB, and the result shows that the extracted signal is similar to the speech signal of the auto-driver, the error is relatively small, and the suppression to the engine noise is obvious.