A UAV-Based Traffic Monitoring System - Invited Paper

摘要

Traffic monitoring is important in urban areas. Traffic sensing solutions based on static cameras, however, do not offer a flexible, inexpensive solution for short-term traffic studies. To overcome the limitations of static solutions, we propose an aerial traffic monitoring system. It uses unmanned aerial vehicles (UAV) and onboard cameras to capture traffic video, which is sent and processed in the cloud. To validate the proposed system, we implemented a prototype with a quadcopter, an onboard camera with video and data processing algorithms, and a web application. The video processing module includes a vehicle detection stage based on the Haar cascade model and a frame-by-frame tracking stage. After experimental testing based on videos collected by the UAV, we conclude that the designed system can monitor traffic with high accuracy and flexibility.