19 May 2015 Change detection on UGV patrols with respect to a reference tour using VIS imagery
Author Affiliations +
Abstract
Autonomous driving robots (UGVs, Unmanned Ground Vehicles) equipped with visual-optical (VIS) cameras offer a high potential to automatically detect suspicious occurrences and dangerous or threatening situations on patrol. In order to explore this potential, the scene of interest is recorded first on a reference tour representing the 'everything okay' situation. On further patrols changes are detected with respect to the reference in a two step processing scheme. In the first step, an image retrieval is done to find the reference images that are closest to the current camera image on patrol. This is done efficiently based on precalculated image-to-image registrations of the reference by optimizing image overlap in a local reference search (after a global search when that is needed). In the second step, a robust spatio-temporal change detection is performed that widely compensates 3-D parallax according to variations of the camera position. Various results document the performance of the presented approach.
© (2015) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Thomas Müller, Thomas Müller, } "Change detection on UGV patrols with respect to a reference tour using VIS imagery", Proc. SPIE 9460, Airborne Intelligence, Surveillance, Reconnaissance (ISR) Systems and Applications XII, 94600I (19 May 2015); doi: 10.1117/12.2182663; https://doi.org/10.1117/12.2182663
PROCEEDINGS
12 PAGES


SHARE
Back to Top