TY - GEN
T1 - Vision-based localization using active scope camera - Accuracy evaluation for structure from motion in disaster environment
AU - Ishikura, Michihisa
AU - Takeuchi, Eijiro
AU - Konyo, Masashi
AU - Tadokoro, Satoshi
PY - 2010/12/1
Y1 - 2010/12/1
N2 - This paper presents the evaluation results for conventional methods that can be used for vision-based localization. An Active Scope Camera is a very thin snake robot and can be used as a rescue robot for search and rescue missions. Self-position estimation of the Active Scope Camera is important for efficient search. Nevertheless, using sensors for this purpose hinders the movement and maneuvering of the camera through narrow gaps, because sensors are very big and heavy for the Active Scope Camera. Vision-based localization using a fish-eye camera is suitable technique for self-position estimation. However, the images obtained using the Active Scope Camera are not of good quality. The material of objects found in disaster environments and overexposure by light-emitting diodes embedded at the camera tip affects the matching of feature points. In this paper, properties of images of disaster sites obtained using the Active Scope Camera and the accuracy evaluation of vision-based localization are described.
AB - This paper presents the evaluation results for conventional methods that can be used for vision-based localization. An Active Scope Camera is a very thin snake robot and can be used as a rescue robot for search and rescue missions. Self-position estimation of the Active Scope Camera is important for efficient search. Nevertheless, using sensors for this purpose hinders the movement and maneuvering of the camera through narrow gaps, because sensors are very big and heavy for the Active Scope Camera. Vision-based localization using a fish-eye camera is suitable technique for self-position estimation. However, the images obtained using the Active Scope Camera are not of good quality. The material of objects found in disaster environments and overexposure by light-emitting diodes embedded at the camera tip affects the matching of feature points. In this paper, properties of images of disaster sites obtained using the Active Scope Camera and the accuracy evaluation of vision-based localization are described.
UR - http://www.scopus.com/inward/record.url?scp=79952799498&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=79952799498&partnerID=8YFLogxK
U2 - 10.1109/SII.2010.5708296
DO - 10.1109/SII.2010.5708296
M3 - Conference contribution
AN - SCOPUS:79952799498
SN - 9781424493159
T3 - 2010 IEEE/SICE International Symposium on System Integration: SI International 2010 - The 3rd Symposium on System Integration, SII 2010, Proceedings
SP - 25
EP - 30
BT - 2010 IEEE/SICE International Symposium on System Integration
T2 - 3rd International Symposium on System Integration, SII 2010
Y2 - 21 December 2010 through 22 December 2010
ER -