Vehicle Speed Estimation and Traffic Tracking System Using Machine Learning

Abstract

In this work, we deploy a real-time method for classifying vehicles and estimating their speeds using footage captured by traffic cameras along motorways. Basic techniques in traffic analysis include the forecasting of traffic flows, the discovery of anomalies, the re-identification of vehicles, and the tracking of moving vehicles. One of the most actively studied areas of these applications is traffic flow prediction, often known as vehicle speed estimation. In this work, we estimate vehicle speeds within classes using feature tracking and neighbor discovery techniques.