机动车声信号能量特征的小波包提取

Energy Feature Extraction of Vehicle Acoustic Signal by Wavelet Packet

针对机动车车型识别中声信号非平稳且易受噪声干扰的问题,提出了一种有效的声信号特征提取方法。利用小波包分析技术对声信号的能量分布进行研究,以德比契斯(Daubechies)小波为基函数对目标声信号进行小波包变换。基于获取的不同频带能量分布状态给出了机动车车型的特征判据,并对该判据的有效性给予了分析。试验结果表明基于小波包分析的机动车声信号特征提取方法是有效的。

In view of the situation that acoustic signal of road-running vehicles is non-stationary and easily drawn in backgroundnoise, a novel approach of feature extraction of the acoustic signal is presented in the paper. Wavelet packet technique is applied to analyze its energy distribution with Daubechies wavelet used as basic function. A feature criterion of vehicle type is obtained based on the energy distribution in each frequency band, and then its validity is analyzed. Experimental results showed that the proposed method was feasible and effective.