Energy and Width Features-Based SVM for Vehicles Classification Using Low Power Consumption Radar

Abstract

In this paper, we propose a novel features-based support vector machines (SVM) vehicles classification approach using low power consumption radar. This new approach exploits the energy and width features, which show different characteristics in two types of vehicles, big car and small car. The proposed approach firstly detects and separates the target mainly using image feature extraction algorithm including grey processing, median filtering, binarization, region growing, filtering small-area blocks and morphological processing. Then coherent accumulation is used to improve the signal-noise ratio and the range unit of target is determined by the endpoint detection algorithm based on dual-threshold. Finally, the energy and width features of each type of car are extracted and the support vector machines (SVM) classifier is applied. The experimental results show that the accuracy of the proposed approach can achieve 95% above.