Paper
1 November 2024 Self-localization method based on visual odometry of autonomous mobile robot for remote-sensing in outdoor underground infrastructure facilities
Author Affiliations +
Abstract
A visual odometry (VO) method is proposed that uses road surface motion images for self-localization by mobile robots. Our proposed method utilizes the feature points obtained from the road surface motion images with the RGB-D camera. Experiments with video images of sub-millimeter resolution show that the VO method is capable of self-position estimation within 1% error. Results also show that the error is corrected by using our proposed algorithm and that the position and posture of the mobile robot in the two-dimensional plane can be self-estimated.
(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.
Masato Mizukami, Yoshikazu Ebina, Takuya Kosakai, and Shoji Mochizuki "Self-localization method based on visual odometry of autonomous mobile robot for remote-sensing in outdoor underground infrastructure facilities", Proc. SPIE 13198, Remote Sensing Technologies and Applications in Urban Environments IX, 1319808 (1 November 2024); https://doi.org/10.1117/12.3028644
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Cameras

Mobile robots

Augmented reality

Autoregressive models

Error analysis

Roads

Visualization

Back to Top