Skip to Main content Skip to Navigation
Conference papers

Line-based Monocular Graph SLAM

Abstract : This paper presents a new line based 6-DOF monocular algorithm that uses the iSAM2, a point-based Graph SLAM approach. We extend iSAM2 to minimize the reprojection error of the line features to solve the line-based SLAM problem. A specific line representation is exploited that combines the Plucker Coordinates and the Cayley representation. The Plucker Coordinates are used for the 3D line projection function, and the Cayley representation helps to update the lines parameters during the non-linear optimization process. An undelayed initialization method with inverse depth parameters is also adopted. Both simulation and real experiments are carried out showing that the approach achieves high accuracy and consistency, and outperforms EKF-based SLAM method.
Complete list of metadatas

Cited literature [21 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01678979
Contributor : Vincent Fremont <>
Submitted on : Tuesday, January 9, 2018 - 3:41:54 PM
Last modification on : Wednesday, May 6, 2020 - 4:18:12 PM
Long-term archiving on: : Wednesday, May 23, 2018 - 5:46:35 PM

File

MFI17_0027_FI.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01678979, version 1

Citation

Dong Ruifang, Vincent Frémont, Simon Lacroix, Isabelle Fantoni, Liu Changan. Line-based Monocular Graph SLAM. IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI 2017), Nov 2017, Daegu, South Korea. pp.494-500. ⟨hal-01678979⟩

Share

Metrics

Record views

595

Files downloads

1082