User: Guest  Login
Title:

Manually annotated and curated dataset of diverse weed species in maize and sorghum for computer vision

Document type:
Forschungsdaten
Publication date:
27.10.2023
Responsible:
Grimm, Dominik G
Authors:
Genze,Nikita (1; 2); Vahl, Wouter K. (4); Groth, Jennifer (4); Wirth, Maximilian (1; 2); Grieb, Michael (5); Grimm, Dominik G. (1; 2; 3)
Author affiliation:
1. Technical University of Munich, TUM Campus Straubing for Biotechnology and Sustainability, Bioinformatics, Schulgasse 22, 94315 Straubing, Germany
2. Weihenstephan-Triesdorf University of Applied Sciences, Bioinformatics, Petersgasse 18, 94315 Straubing, Germany
3. Technical University of Munich, TUM School of Computation, Information and Technology (CIT), Boltzmannstr. 3, 85748 Garching, Germany
4. Institute for Crop Science and Plant Breeding, Bavarian State Research Center for Agriculture, Am Gereuth 6, 85354 Freising, Germany
5. Technology and Support Centre in the Centre of Excellence for Renewable Resources (TFZ), Schulgasse 18, 94315 Straubing, Germany
Publisher:
TUM
Identifier:
doi:10.14459/2023mp1717366
End date of data production:
13.12.2021
Subject area:
BIO Biowissenschaften; DAT Datenverarbeitung, Informatik; LAN Landbauwissenschaft; NAT Naturwissenschaften (allgemein)
Resource type:
Abbildungen von Objekten / image of objects; Textdokumente / text documents
Data type:
Bilder / images; Tabellen / tables
Other data type:
csv
Description:
To generate a dataset consisting of high quality images that capture the initial growth dynamics of individual plants of several weed species, a greenhouse experiment was performed at the Moving Fields facility of the LfL in Freising, Germany. This experiment took place from 16.06.2021 - 13.12.2021. The plant species were selected as weed species common to fields of sorghum grown in Germany. A Scanalyzer 3D imaging cabin of the Moving Fields facility was used for generating high-quality images....     »
Method of data assessment:
The plants were grown in plots that were watered and photographed automatically on at least a daily basis from the day of sowing till harvest, which took place at around shooting. Afterwards plants in a plot was annotated with bounding boxes using the open souce software CVAT (cvat.ai) throughout the growth, leading to a timeseries of images from sprouting to harvest. Not each plant in a plot was annotated, leading to plots that were fully annotated (usable in object detection and tracking) and partly annotated - which can be used for classification and pre-training. Additionally, we provide some images of trays that were not annotated.
Finally, we provide instance segmentation masks for a total of >2.000 plant samples of 14 different species.
Links:
This dataset relates to the publication: https://www.nature.com/articles/s41597-024-02945-6
Key words:
Weed Detection; Computer Vision; Object Detection; Weed Classification; Segmentation; Time-series
Technical remarks:
View and download (928 GB total, 421 Files)
The data server also offers downloads with FTP
The data server also offers downloads with rsync (password m1717366):
rsync rsync://m1717366@dataserv.ub.tum.de/m1717366/
Language:
en
Rights:
by, http://creativecommons.org/licenses/by/4.0
 BibTeX