Nikos Papanikolopoulos, Professor, Computer Science and Engineering
This research is investigating the use of multiple cameras for monitoring human activities at critical transportation infrastructure sites, leveraging work performed in the researchers' Department of Homeland Security (DHS) project. The methods deal with detecting specified activities and counting humans in crowded scenes. The researchers will further develop methods to automatically detect and spatially estimate an occlusion (common in crowded outdoor scenes) in world coordinates. The algorithms will be tested at the transit stations where the DHS system is currently deployed and will cover activities that are not of interest to the DHS but are of major interest to Metro Transit (e.g., loitering, graffiti, drug dealing). The proposed methods, which focus on human activities, are directly applicable to a wide variety of transportation infrastructure sites.