An Approach of Passive Vehicle Type Recognition by Acoustic Signal Based on SVM

Abstract

An approach of power spectrum estimation is utilized to extract the feature vectors from acoustic signal radiated from different types of moving vehicles. A method of feature selection based on principal component analysis (PCA) is proposed to reconstruct effective feature vectors via dimension reduction. The classification of three typical targets is achieved by supported vector machine (SVM). Experiment results show that the approach presented in the paper for automatic recognition of vehicle type is effective.