Real-Time, Multiple, Pan/Tilt/Zoom, Computer Vision Tracking, and 3D Position Estimating System for Small Unmanned Aircraft System Metrology.

摘要

The study of the flight kinematics of Unmanned Aircraft Systems (UASs) continues to be an important field of research for understanding and developing state-of-the-art nano/micro air vehicles. Analyzing various flapping-wing UASs using photogrammetry shows potential given the ability to use images for constructing a three-dimensional (3D) representation of a given object. Development of a photogrammetry/metrology system using Computer Vision (CV) tracking and 3D point extraction would provide an avenue for collecting free-flight data crucial in understanding the aerodynamic and aeroelastic behavior of flapping wing UASs; thus far not achieved. Free-flight data would provide a link between empirical and theoretical development thereby increasing knowledge of various flight profiles of flapping-wing UASs. The focus of this work is to develop a scalable system capable of real-time tracking, zooming, and 3D position estimation of a UAS using multiple cameras.;Real-time, state-of-the-art photogrammetry systems use retro-reflective markers to obtain object poses and/or positions over time. Computer Vision (CV) methods capable of performing photogrammetry tasks provide the potential for non-contact measurements that avoid altering the test subject. Related work includes using Pan/Tilt/Zoom (PTZ) cameras in combination with or without static cameras; however, little has been done in the area of metrology and multiple PTZ cameras. Contributions include the development of a real-time CV-tracking algorithm for PTZ cameras, using optical flow and known camera motion, in order to capture a moving object. The developed implementation is versatile, allowing the ability to test other CV methods with a PTZ system and known camera motion. Further, digital zoom is applied to each camera to simulate PTZ cameras for tracking purposes. Utilizing known camera poses, the UAS's 3D position is estimated and focal lengths are calculated for filling the image to a desired amount. This system is tested against truth data obtained using a commercial photogrammetry system. Automatically capturing high- resolution, high frame-rate images of small, free-flight flappers is obtainable using the system developed in this work.