Video Analytics for Vision Zero III refines the tools and work conducted in earlier Video Analytics for Vision Zero tasks by testing the use of automated object recognition/tracking in video data streams to support the assessment of pedestrian safety using existing traffic monitoring cameras. The specific tasks pursued in the scope include: 1) refine the video recording workflow; 2) video recording and analysis at the Lamar/Rundberg and Lamar/Payton Gin locations; 3) analysis of bus stop activities and crossing event detection; 4) validation of crossing detection with PHB device activations; and 5) creating a web interface for interactive video reviewing. The two use cases examined in work task 2 illustrate how a scalable tool for automated analysis of data collected from monocular traffic cameras can allow agencies to leverage existing infrastructure in analysis and mitigation of pedestrian safety concerns.
Read the final report here.