Statistical profile generation for traffic monitoring using real-time UAV based video data

Abstract

The eye-in-the-sky alternative to collecting real-time temporal/spatial data using small unmanned helicopters is proposed to: monitor traffic, evaluate and assess traffic patterns and provide accurate vehicle counts. Collected real-time visual data are converted to traffic statistical profiles and used as continuously updated inputs to existing traffic simulation models improving calibration, accuracy (in terms of variable parameter values) and future traffic predictions. Functionality of simulation models is enhanced, and reliability is improved The proposed approach offers significant advantages over conventional methods where historical and outdated data is used to run poorly calibrated traffic simulation models.