REAL-TIME VISUAL ODOMETRY FOR GROUND MOVING ROBOTS USING GPUS

Michael Schweitzer, Alois Unterholzner, Hans-Joachim Wuensche

2010

Abstract

This paper introduces a novel visual odometry framework for ground moving robots. Recent work showed that assuming non-holonomic motion can simplify the ego motion estimation task to one yaw and one scale parameter. Furthermore, a very efficient way of computing image frame to frame correspondences for those robots was presented by skipping rotational invariance and optimizing keypoint extraction and matching for massive parallelism on a GPU. Here, we combine both contributions to a closed framework. Long term correpondences are preserved, classified and stablized by motion prediction, building up and keeping a trusted map of depth-registered keypoints. We also allow other ground moving objects. From this map, the ego motion is infered, extended by constrained rotational perturbations in pitch and roll. A persistent focus is on keeping algorithms suitable for parallelization and thus achieving up to one hundred frames per second. Experiments are carried out to compare against ground-truth given by DGPS and IMU data.

Paper Citation

in Harvard Style

Schweitzer M., Unterholzner A. and Wuensche H. (2010). REAL-TIME VISUAL ODOMETRY FOR GROUND MOVING ROBOTS USING GPUS . In Proceedings of the International Conference on Computer Vision Theory and Applications - Volume 1: VISAPP, (VISIGRAPP 2010) ISBN 978-989-674-028-3, pages 20-27. DOI: 10.5220/0002821200200027

in Bibtex Style

@conference{visapp10,
author={Michael Schweitzer and Alois Unterholzner and Hans-Joachim Wuensche},
title={REAL-TIME VISUAL ODOMETRY FOR GROUND MOVING ROBOTS USING GPUS},
booktitle={Proceedings of the International Conference on Computer Vision Theory and Applications - Volume 1: VISAPP, (VISIGRAPP 2010)},
year={2010},
pages={20-27},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0002821200200027},
isbn={978-989-674-028-3},
}

in EndNote Style

TY - CONF
JO - Proceedings of the International Conference on Computer Vision Theory and Applications - Volume 1: VISAPP, (VISIGRAPP 2010)
TI - REAL-TIME VISUAL ODOMETRY FOR GROUND MOVING ROBOTS USING GPUS
SN - 978-989-674-028-3
AU - Schweitzer M.
AU - Unterholzner A.
AU - Wuensche H.
PY - 2010
SP - 20
EP - 27
DO - 10.5220/0002821200200027