基于经验模态分解和支持向量机的车型声频识别

Vehicle recognition by acoustic signals based on EMD and SVM

车型自动识别是智能交通系统的重要组成部分。针对现有车型识别存在的问题,提出利用经验模态分解和支持向量机的车型声频识别方法。将车辆行驶的声音信号进行分解,以分解不同模态的能量作为特征向量,并以此作为训练样本对支持向量机构成的车型识别器进行训练,通过对小汽车和卡车的声音信号处理结果表明:利用车辆声音信号能够正确识别不同的车型,识别准确率达95%,是车型识别的有效方法。

Automatic vehicle recognition is an important part of the Intelligent Transportation Systems. For the problem of automatic vehicle recognition, a recognition method is proposed based on EMD and support vector machine. Vehicle acoustic signals are decomposed with the EMD, and the powers in different intrinsic mode functions are regarded as the different vehicle eigenvectors and to be uses as training samples of the SVM vehicle classifier. By the processing of car and truck acoustic signals, the result shows that various vehicles can be identified using the vehicle sound signals, the recognition rate is 95%, and it is an effective method for automatic vehicle recognition.