English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In recent years, unsupervised deep learning approaches have receivedsignificant attention to estimate the depth and visual odometry (VO) fromunlabelled monocular image sequences. However, their performance is limited inchallenging environments due to perceptual degradation, occlusions and rapidmotions. Moreover, the existing unsupervised methods suffer from the lack ofscale-consistency constraints across frames, which causes that the VOestimators fail to provide persistent trajectories over long sequences. In thisstudy, we propose an unsupervised monocular deep VO framework that predictssix-degrees-of-freedom pose camera motion and depth map of the scene fromunlabelled RGB image sequences. We provide detailed quantitative andqualitative evaluations of the proposed framework on a) a challenging datasetcollected during the DARPA Subterranean challenge; and b) the benchmark KITTIand Cityscapes datasets. The proposed approach outperforms both traditional andstate-of-the-art unsupervised deep VO methods providing better results for bothpose estimation and depth recovery. The presented approach is part of thesolution used by the COSTAR team participating at the DARPA SubterraneanChallenge.

Unsupervised Deep Persistent Monocular Visual Odometry and Depth Estimation in Extreme Environments