Public and Private Vehicle Quantification and Classification using Vehicle Detection and Recognition

Abstract

Traffic congestion in the Philippines is diverse consisting public and private vehicles. One approach of this is to design a system that can count, detect, recognize, and classify public and private vehicles from a surveillance video. This research introduces the development of the said system, to be used as a statistical data for implementing traffic rules. The dataset the researchers used consists of 13,600 images: 10,880 images for training and 2,720 images for testing. These were obtained from a gas station video source, for the vehicles frequently passing though in a gas station. The researchers used an algorithm called Convolutional Neural Network for the detecting and classifying vehicles.