Premium
A Benchmark for Outdoor Vision SLAM Systems
Author(s) -
Abdallah Samer M.,
Asmar Daniel C.,
Zelek John S.
Publication year - 2007
Publication title -
journal of field robotics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.152
H-Index - 96
eISSN - 1556-4967
pISSN - 1556-4959
DOI - 10.1002/rob.20180
Subject(s) - simultaneous localization and mapping , benchmark (surveying) , artificial intelligence , computer vision , landmark , computer science , dead reckoning , inertial measurement unit , robot , robustness (evolution) , machine vision , mobile robot , global positioning system , geography , cartography , telecommunications , biochemistry , chemistry , gene
Simultaneous localization and mapping (SLAM) is a viable solution to autonomous robot navigation in outdoor settings when global positioning systems are unavailable or unreliable. While the traditional exteroceptive sensor for outdoor SLAM is a laser, cameras have also been proposed due to their low power consumption, low price, high bandwidth of information, and superior landmark segmentation capabilities. All outdoor Vision SLAM systems developed to date are implemented on different platforms, in different settings, using different dead‐reckoning sensors; a fact which makes it difficult to compare them and assess the state of the art of Vision SLAM. The contribution of this paper is in developing an infrastructure for a benchmark upon which past and future Vision SLAM system can be compared. This proposed benchmark is validated by testing its datasets on a Vision‐Inertial SLAM system. © 2007 Wiley Periodicals, Inc.