Benutzer: Gast  Login
Titel:

Dataset for Instance Segmentation of Deformable Linear Objects

Dokumenttyp:
Forschungsdaten
Veröffentlichungsdatum:
14.12.2022
Verantwortlich:
Dirr, Jonas
Autorinnen / Autoren:
Dirr, Jonas; Yao, Jiajun; Siepmann, Andre; Gebauer, Daniel; Daub, Rüdiger
Institutionszugehörigkeit:
TUM
Herausgeber:
TUM
Identifikator:
doi:10.14459/2022mp1690303
Enddatum der Datenerzeugung:
12.08.2022
Fachgebiet:
DAT Datenverarbeitung, Informatik; MAS Maschinenbau
Quellen der Daten:
Abbildungen von Objekten / image of objects
Datentyp:
Bilder / images
Beschreibung:
This dataset contains images of deformable linear objects (DLOs) in bin picking applications. The images show cables from control cabinet construction, which are provided in small load carriers. The dataset is divided into three scenarios, which differ in terms of cable arrangement:

>S1 Individual provision: One single DLO is provided in the small load carrier.
>S2 Semi-structured provision: Multiple DLOs are provided in the small load carrier, but they do not overlap.
>S3 Unstructured provision: In a small load carrier, multiple DLOs are provided and may overlap or cross each other.

For each scenario, one subfolder with about 80 real-world images in JPG format is provided. The manually annotated instance masks are provided in COCO format as JSON files. In addition, the instance masks of the cables are visualized as JPG images.
Links:
This dataset relates to the publication: https://doi.org/10.3390/s23063013
Schlagworte:
deformable linear object; cable; wire; instance segmentation; bin picking, object detection
Technische Hinweise:
View and download (45 MB total, 493 Files)
The data server also offers downloads with FTP
The data server also offers downloads with rsync (password m1690303):
rsync rsync://m1690303@dataserv.ub.tum.de/m1690303/
Sprache:
en
Rechte:
by-sa, http://creativecommons.org/licenses/by-sa/4.0
 BibTeX