Performance Evaluation of a Low Cost Vision-Based Traffic Flow Monitoring System

Abstract—Traffic flow monitoring systems aim to provide accurate, complete and timely data for effective management of congestion. Conventional traffic flow monitoring systems are not suitable for deployment in crowed cities in the developing-world, due to the often chaotic nature of traffic there and typically limited budgets. In this paper we present the design and implementation of a low cost vision-based traffic flow monitoring system for such contexts using a mobile phone camera as a unit for data capture and transmission. We describe its architecture and hardware platform as well as the procedure used in traffic flow speed estimation. We also evaluate the system with performance using field experiments in Kampala, Uganda. From the experiments, we observe that the system provides a feasible method of continuously monitoring traffic congestion while reducing deployment costs drastically compared to other technologies in current use. The system also performs well according to other criteria such as ability to operate without maintenance. I