PolaBot (2019)

Available

Dataset

Licence : CC BY-NC-SA doi:10.25666/DATAUBFC-2022-05-13-02

Access to data

Data creators : Ralph Seulin ^[1], Raphael Duverne ^[1], Marc Blanchon ^[1], Olivier Morel ^[1]

Related person : Désiré Sidibé ^[2], Fabrice Mériaudeau ^[1]

[1] : Imagerie et Vision Artificielle (UR 7535) (Université de Bourgogne)

[2] : Université d'Évry-Val-d'Essonne

Description :

PolaBot is a dataset dedicated to the polarimetric imaging of outdoor scenes.
Outdoor scenes include challenging surfaces to segment and detect : highly reflective areas such as water and windows.
This dataset is designed for research and development fields such as mobile robotics, autonomous cars and navigation, ...
In addition, this collection of information will allow a strong and efficient benchmark, giving the opportunity to compare standard modality to the polarimetry for the exact same scenes and application.

Disciplines :

computer science, software engineering (engineering science), imaging science & photographic technology (engineering science), robotics (engineering science), transportation science & technology (engineering science)

Keywords :

computer vision (Electrotechnique), perception applied to robotics, polarimetric imaging, polarimetry, signal processing (Electrotechnique)

General metadata

Data acquisition date : from 2018 to 2019

Data acquisition methods :

Experimental data

Update periodicity : no update

Language : French (fra)

Formats : image/png, text/plain

Audience : Research

Publications :

M. Blanchon, D. Sidibé, O. Morel, R. Seulin, D. Braun and F. Meriaudeau, "P2D: a self-supervised method for depth estimation from polarimetry," 2020 25th International Conference on Pattern Recognition (ICPR), 2021, pp. 7357-7364 (doi:10.1109/ICPR48806.2021.9412441)
M. Blanchon, O. Morel, F. Meriaudeau, R. Seulin and D. Sidibé, "Polarimetric image augmentation," 2020 25th International Conference on Pattern Recognition (ICPR), 2021, pp. 7365-7371 (doi:10.1109/ICPR48806.2021.9412133)
M. Blanchon, “Polarization based urban scenes understanding” (manuscrit de thèse), 2021 (hal:tel-03469970v1)

Collection :

Données de thèse

Publisher : Imagerie et Vision Artificielle (UR 7535)

Projects and funders :

Thèse de doctorat en instrumentation et informatique de l'image - Marc Blanchon
- Contrat doctoral de droit public - projet ANR-15-CE22-0009 (French National Agency for Research)
Polarimetric Vision Applied to Robotics Navigation - ViPeR
- ANR-15-CE22-0009 (French National Agency for Research)

Additional information :

Each acquisition consists of three images: the RGB color image, the raw image from the polarimetric camera and the annotated image.
Data collected as part of the thesis in instrumentation and computer vision by Marc Blanchon, co-supervised by Fabrice Mériaudeau (professor) and Olivier Morel (associated professor), ImViA, university of Burgundy, and Désiré Sidibé (professor) university of Evry.

Data

	1 file
		PolaBot.zip Published : 21/03/2024 12:03 Size : 583.70 MB Continued Description : PolaBot is a dataset dedicated to the polarimetric imaging of outdoor scenes. Outdoor scenes include challenging surfaces to segment and detect : highly reflective areas such as water and windows. This dataset is designed for research and development fields such as mobile robotics, autonomous cars and navigation, ... In addition, this collection of information will allow a strong and efficient benchmark, giving the opportunity to compare standard modality to the polarimetry for the exact same scenes and application.

1 file

PolaBot.zip

Published : 21/03/2024 12:03 Size : 583.70 MB

Continued

Description : PolaBot is a dataset dedicated to the polarimetric imaging of outdoor scenes.
Outdoor scenes include challenging surfaces to segment and detect : highly reflective areas such as water and windows.
This dataset is designed for research and development fields such as mobile robotics, autonomous cars and navigation, ...
In addition, this collection of information will allow a strong and efficient benchmark, giving the opportunity to compare standard modality to the polarimetry for the exact same scenes and application.