Automating vehicular traffic estimation using optical flow analysis
Abstract
We present an algorithm for movement speed estimation, and demonstrate the errors that may occur when cameras are deployed in less-than-ideal conditions, such as when there is much occlusion in the video, or the video is taken with camera phones or from security feeds. We find speed underestimation occurs when static occlusions are present, and overestimation with moving occlusions.
Downloads
Issue
International Year of Light and light-based technologies
3-6 June 2015, University of Northern Philippines, Vigan City