Accurate 6D Object Pose Estimation and Refinement in Cluttered Scenes

Yixiang Jin, John Rossiter, Sandor Veres

2021

Abstract

Estimating the 6D pose of objects is an essential part of a robot’s ability to perceive their environment. This paper proposes a method for detecting a known object and estimating its 6D pose from a single RGB image. Unlike most of the state-of-the-art methods that deploy PnP algorithms for estimating 6D pose, the method here can output the 6D pose in one step. In order to obtain estimation accuracy that is comparable to RGB-D based methods, an efficient refinement algorithm, called contour alignment (CA), is presented; this can increase the predicted 6D pose accuracy significantly. We evaluate the new method in two widely used benchmarks, LINEMOD for single object pose estimation and Occlusion-LINEMOD for multiple objects pose estimation. The experiments show that the proposed method surpasses other state-of-the-art prediction approaches.

Paper Citation

in Harvard Style

Jin Y., Rossiter J. and Veres S. (2021). Accurate 6D Object Pose Estimation and Refinement in Cluttered Scenes. In Proceedings of the 2nd International Conference on Robotics, Computer Vision and Intelligent Systems - Volume 1: ROBOVIS, ISBN 978-989-758-537-1, pages 31-39. DOI: 10.5220/0010654500003061

in Bibtex Style

@conference{robovis21,
author={Yixiang Jin and John Rossiter and Sandor Veres},
title={Accurate 6D Object Pose Estimation and Refinement in Cluttered Scenes},
booktitle={Proceedings of the 2nd International Conference on Robotics, Computer Vision and Intelligent Systems - Volume 1: ROBOVIS,},
year={2021},
pages={31-39},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010654500003061},
isbn={978-989-758-537-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Robotics, Computer Vision and Intelligent Systems - Volume 1: ROBOVIS,
TI - Accurate 6D Object Pose Estimation and Refinement in Cluttered Scenes
SN - 978-989-758-537-1
AU - Jin Y.
AU - Rossiter J.
AU - Veres S.
PY - 2021
SP - 31
EP - 39
DO - 10.5220/0010654500003061