Vision-Aided Navigation for Aerial Platforms

Leonid Naimark, Helen Webb and Tao Wang

Abstract:	We combine altimeter, IMU-provided orientation, and along-path to reference imagery registration to provide airborne platform navigation solutions in GPS-denied environments, at an accuracy and precision primarily gated by the quality of locally available digital elevation maps (DEM) and reference imagery. Our image-to-image registration approach, based on an enhancement to phase correlation which greatly increases its efficacy, significantly outperforms corner-based methods. Sensor orientation, required for image projection, is provided by even inexpensive IMUs, which experience limited error indefinitely, independent of the amount of accumulated drift in position. When feature-rich imagery becomes available (e.g. after transitioning from over-water to over-land flight), a limited iteration over rotation parameters of the vehicle allows the recapture of the pose. For standard errors in altitude, IMU orientation, and reference DEMs and imagery, our approach typically provides combined navigation accuracy to within several meters, often within 1 meter - thus in many cases outperforming GPS.
Published in:	Proceedings of the ION 2017 Pacific PNT Meeting May 1 - 4, 2017 Marriott Waikiki Beach Resort & Spa Honolulu, Hawaii
Pages:	70 - 76
Cite this article:	Naimark, Leonid, Webb, Helen, Wang, Tao, "Vision-Aided Navigation for Aerial Platforms," Proceedings of the ION 2017 Pacific PNT Meeting, Honolulu, Hawaii, May 2017, pp. 70-76. https://doi.org/10.33012/2017.15051
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In