Robust Direct Visual Odometry Estimation for a Monocular Camera Under Rotations

Lakshmi, A.; AGJ, Faheema; Deodhare, Dipti

doi:10.1109/lra.2017.2755070

Cited by 2 publications

(3 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We name this new rule Rule1. Then, we define Rule2 with Step 2 initialized with λ = λ * = 1 (or smaller) and λ optimized for more coherence with respect to Step 1, more freedom and to use the exact same control law (3). For all rules, the switch from Step 1 to Step 2 occurs after a fixed number of iterations n it ∈ N * to ease the comparisons.…”

Section: The Two Stages Strategy: New Rulesmentioning

confidence: 99%

“…Direct image alignment [1] and direct visual servoing [2] (DVS) have significantly progressed during the last decade in their respective communities, namely computer vision and robotics. The best known direct approaches concern visual odometry [3] and visual Simultaneous Localization And Mapping [4] (SLAM). For a while, direct approaches were known to save time by avoiding features processing and to be of high accuracy whereas suffering of a narrow convergence domain [5].…”

Section: Introductionmentioning

confidence: 99%

“…with geometric algorithms of https://www.cloudcompare.org3 We share our implementation of image transformation on repository https://github.com/PerceptionRobotique/equi2omni…”

mentioning

confidence: 99%

See 2 more Smart Citations

Direct 3D model-based tracking in omnidirectional images robust to large inter-frame motion

Guerbas

Crombez²,

Caron³

et al. 2021

2021 20th International Conference on Advanced Robotics (ICAR)

View full text Add to dashboard Cite

This paper tackles direct 3D model-based pose tracking. It considers the Photometric Gaussian Mixtures (PGM) transform of omnidirectional images as direct features. The contributions include an adaptation of the pose optimization to omnidirectional cameras and a rethink of the initialization and optimization rules of the PGM extent. These enhancements produce a giant leap in the convergence domain width. Application to images acquired onboard a mobile robot within an urban environment described by a large 3D colored point cloud shows significant robustness to large interframe motion, compared to approaches that directly use pixel brightness as direct features.

show abstract

Section: The Two Stages Strategy: New Rulesmentioning

confidence: 99%