Traffic incident automatic detection and recognition play an important role in traffic management and controlling. For the problems of automatic traffic incident detection and recognition, a method is proposed based on wavelet decomposition and support vector machine (SVM) with vehicle acoustic signals. Vehicle acoustic signals are decomposed with the wavelet analysis, and the power in different frequency is regarded as the different incident eigenvectors and also as training samples of SVM traffic incident classifier. By the processing of normal driving, braking and crash incident acoustic signals, the result shows that the proposed method is an effective method for automatic traffic incident detection and recognition.
Traffic Incident Acoustic Recognition Based on Wavelet Decomposition and SVM
Ninth International Conference of Chinese Transportation Professionals (ICCTP) ; 2009 ; Harbin, China
ICCTP 2009 ; 1-6
2009-07-23
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Traffic Incident Acoustic Recognition Based on Wavelet Decomposition and SVM
British Library Conference Proceedings | 2009
|Traffic Incident Detection Based on Wavelet and Neural Network
British Library Conference Proceedings | 2009
|A Detection Method for Traffic Incident Using Wavelet Analysis
British Library Conference Proceedings | 2001
|Incident Detection Algorithm using Wavelet Energy Representation of Traffic Patterns
British Library Online Contents | 2002
|