3 April 2008 Global image registration using shape space tracking
Author Affiliations +
Abstract
In airborne UAV operations, it is desirable to havemultiple UAVs operate in a cooperationmode tomaximize the use of resources, as well as take unique advantages of increased views frommultiple positions separated in time and space. This capability requires image registration across potentially widely different views that are taken at different places or time. The resulting images also vary in their noise contents, resolutions, and projection deformation. In this paper, we characterize the performance enhancement between single view from a single UAV and multiple views by multiple UAVs using ROC curves. In addition, we propose a computational framework to facilitate an efficient and accurate registration across images with varied resolutions, noise levels, and projective deformation. We implement various stages in this framework and demonstrate the promising results using low to medium resolution images with synthetically-generated flying path and camera poses.
© (2008) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Liangyin Yu, Liangyin Yu, Jóse Molineros, Jóse Molineros, Venkataraman Sundareswaran, Venkataraman Sundareswaran, } "Global image registration using shape space tracking", Proc. SPIE 6981, Defense Transformation and Net-Centric Systems 2008, 69810G (3 April 2008); doi: 10.1117/12.779409; https://doi.org/10.1117/12.779409
PROCEEDINGS
11 PAGES


SHARE
Back to Top