Semi-supervised Learning for Image Alignment in Teach and Repeat Navigation

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F22%3A00361048" target="_blank" >RIV/68407700:21230/22:00361048 - isvavai.cz</a>

Výsledek na webu

<a href="https://doi.org/10.1145/3477314.3507045" target="_blank" >https://doi.org/10.1145/3477314.3507045</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1145/3477314.3507045" target="_blank" >10.1145/3477314.3507045</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Semi-supervised Learning for Image Alignment in Teach and Repeat Navigation

Popis výsledku v původním jazyce

Visual teach and repeat navigation (VT&R) is a framework that enables mobile robots to traverse previously learned paths. In principle, it relies on computer vision techniques that can compare the camera's current view to a model based on the images captured during the teaching phase. However, these techniques are usually not robust enough when significant changes occur in the environment between the teach and repeat phases. In this paper, we show that contrastive learning methods can learn how the environment changes and improve the robustness of a VT&R framework. We apply a fully convolutional Siamese network to register the images of the teaching and repeat phases. Their horizontal displacement between the images is then used in a visual servoing manner to keep the robot on the intended trajectory. The experiments performed on several datasets containing seasonal variations indicate that our method outperforms state-of-the-art algorithms tailored to the purpose of registering images captured in different seasons.

Název v anglickém jazyce

Semi-supervised Learning for Image Alignment in Teach and Repeat Navigation

Popis výsledku anglicky

Visual teach and repeat navigation (VT&R) is a framework that enables mobile robots to traverse previously learned paths. In principle, it relies on computer vision techniques that can compare the camera's current view to a model based on the images captured during the teaching phase. However, these techniques are usually not robust enough when significant changes occur in the environment between the teach and repeat phases. In this paper, we show that contrastive learning methods can learn how the environment changes and improve the robustness of a VT&R framework. We apply a fully convolutional Siamese network to register the images of the teaching and repeat phases. Their horizontal displacement between the images is then used in a visual servoing manner to keep the robot on the intended trajectory. The experiments performed on several datasets containing seasonal variations indicate that our method outperforms state-of-the-art algorithms tailored to the purpose of registering images captured in different seasons.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/GC20-27034J" target="_blank" >GC20-27034J: Rozšíření prostorových modelů explicitní representací času pro dlouhodobou autonomii mobilních robotů</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2022

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Proceedings of the 37th ACM/SIGAPP Symposium on Applied Computing

ISBN

978-1-4503-8713-2

ISSN

—

e-ISSN

—

Počet stran výsledku

8

Strana od-do

731-738

Název nakladatele

ACM

Místo vydání

New York

Místo konání akce

Virtual

Datum konání akce

25. 4. 2022

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—