Vehicle Classification Based on CCTV Video Recording Using Histogram of Oriented Gradients, Local Binary Patterns, and Hierarchical Multi-SVM

Abstract

Abstract Vehicle classification is a crucial element in a dynamic traffic management system. The system is needed since the number of vehicles gradually increase every year. There are many methods for vehicle classification, one of those is the use of CCTV camera. As it is multifunction and cheaper than other types of camera, the government tend to use the CCTV camera as monitoring equipment of vehicles on the road. Hence, an accurate vehicle classification system from CCTV records is needed for traffic management system as well as other systems. In this study, the method contains four steps: object detection using Gaussian Mixture Model, features extraction using Histogram of Oriented Gradients and Local Binary Patterns, model training and classification using Hierarchical Multi-SVM. The experimental results show that our proposed method works well with the accuracy of 80.28%, precision of 94.27%, recall of 73.79%, and F-measure of 82.76%.