Deeply-Integrated Feature Tracking for Embedded Navigation

Jeff R. Gray and Michael J. Veth

Abstract:	The Air Force Institute of Technology is investigating techniques to improve aircraft navigation using low-cost imaging and iner- tial sensors [1]. Stationary features tracked within the image are used to improve the inertial navigation estimate. Features are tracked using a correspondence search between frames. Previous research investigated aiding these correspondence searches using inertial measurements (i.e., stochastic projection) [2]. While this research demonstrated the bene¯ts of further sensor integration, it still relied on robust feature descriptors (e.g., SIFT or SURF) to obtain a reliable correspondence match in the presence of rotation and scale changes. Unfortunately, these robust feature extraction algorithms are computationally intensive and require signi¯cant resources for real-time operation. Simpler feature extraction algorithms are much more e±cient, but their feature descriptors are not invariant to scale, rotation, or a±ne warping which limits match- ing performance during arbitrary motion. This research uses inertial measurements to predict not only the location of the feature in the next image but also the feature descriptor, resulting in robust correspondence matching with low computational overhead. Navigation experiments using the previous and newly developed ¯lter updates are presented.
Published in:	Proceedings of the 2009 International Technical Meeting of The Institute of Navigation January 26 - 28, 2009 Disney's Paradise Pier Hotel Anaheim, CA
Pages:	1018 - 1025
Cite this article:	Gray, Jeff R., Veth, Michael J., "Deeply-Integrated Feature Tracking for Embedded Navigation," Proceedings of the 2009 International Technical Meeting of The Institute of Navigation, Anaheim, CA, January 2009, pp. 1018-1025.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In