摘要: |
A low-cost aerial platform represents a flexible tool for acquiring high-resolution images for ground areas of interest. The geo-referencing of objects within these images could benefit civil engineers in a variety of research areas including, but not limited to, work zone management, traffic congestion, safety, and environmental impact studies. During the Phase II effort, a Remotely Controlled (R/C) aircraft based remote sensing platform was developed and flight tested at West Virginia University (WVU). Main components of the remote-sensing payload system include a high-resolution digital still camera, a 50 Hz GPS receiver, a low-cost Inertial Navigation System (INS), a down-looking laser range finder, a custom-designed flight data recorder, and a wireless video transmission system. An extensive time-calibration and analysis effort for major measurement instruments was performed to assure that flight data were properly time-aligned. Additionally, an Unscented Kalman Filter (UKF) based 15-state GPS/INS sensor fusion algorithm was developed to estimate the aircraft attitude angles in flight. Based on the added range and orientation information for the camera, the geo-referencing software developed in Phase I effort was enhanced. The performance of the software was evaluated using a set of flight data and the known location of a fixed reference point on the ground. The flight data analysis shows an approximately 7.2 meter mean position estimation error was achieved with estimates from a single aerial image, after a set of lens distortion and the camera orientation corrections. Furthermore, a 0.5 meter position estimation error was achieved with the averaging of 15 individual estimates. |