Monocular Visual Odometry Using Unsupervised Deep Learning

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper proposes a new monocular visual odometry system based on unsupervised deep learning that can simultaneously estimate the relative 6-DoF (degree of freedom) poses and scene depth of monocular cameras from sequences of images. It jointly trains depth maps and relative camera poses without ground truth. The depth model uses the encoder-decoder architecture and the pose model uses RCNN (Recurrent Convolutional Neural Network) with attention mechanism innovatively. The system is trained by using stereo images and tested by using monocular images. Experiments show that this unsupervised monocular visual odometry system doesn't suffer from the scale ambiguity of monocular visual odometry and it outperforms some state of the art approaches.

Original languageEnglish
Title of host publicationProceedings - 2019 Chinese Automation Congress, CAC 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages3274-3279
Number of pages6
ISBN (Electronic)9781728140940
DOIs
StatePublished - Nov 2019
Event2019 Chinese Automation Congress, CAC 2019 - Hangzhou, China
Duration: 22 Nov 201924 Nov 2019

Publication series

NameProceedings - 2019 Chinese Automation Congress, CAC 2019

Conference

Conference2019 Chinese Automation Congress, CAC 2019
Country/TerritoryChina
CityHangzhou
Period22/11/1924/11/19

Keywords

  • RCNN
  • attention mechanism
  • depth estimation
  • unsupervised deep learning
  • visual odometry

Fingerprint

Dive into the research topics of 'Monocular Visual Odometry Using Unsupervised Deep Learning'. Together they form a unique fingerprint.

Cite this