Uses of Airborne Imagery for Microscopic Traffic Analysis

Over the past several years, traffic analysts and researchers have been exploring the use of airborne imagery to understand traffic behavior and to analyze traffic speeds, volumes, densities, delays, queue lengths, and other traffic parameters. As part of the National Consortium on Remote Sensing in Transportation - Flows (NCRST-F), researchers at the University of Arizona have been developing software tools to assist in analyzing traffic behavior, at the level of the individual vehicle (or, "microscopic" traffic behavior). To do this, we have gathered digital video segments taken from a helicopter over traffic facilities, using a digital video camera. The video image sequence is then processed using software TRAVIS (Tracking and Registration of Airborne Video Image Sequence). The software registers the video sequence (eliminating the movement of the aircraft from the image), identifies vehicles in the images, and tracks the vehicles through the video image sequence. The vehicle coordinates in the image sequence can then be transformed into relevant traffic measures: location, speed, acceleration and deceleration, and position in terms of its lane on the roadway. With a large set of video segments, the software has been able to identify and track individual vehicles automatically, thereby significantly reducing the need for manual data reduction for some traffic applications. We also discuss the applications of the data from airborne imagery for understanding microscopic traffic flow and for the development and calibration of traffic simulation models.