Parkalytics

Automatic Georeferencing for UAV video

World coordinates for observations in UAV data are valuable for tracking real objects e.g., vehicles. However, robust automatic matching between image and world coordinates remains elusive. Changes in camera pose are common in video captured from UAVs, and can be challenging. This student team will work to build a process for automatically georeferencing video from UAVs, detect and match features for georeferencing, track and update features under pose changes, and output georeferenced video and transforms. The key milestones this student team will work to accomplish are: - a Literature review and data set gathering - A basic prototype that is functional - Field testing and integration -Iteration