BiSID-5k: A Bimodal Image Dataset for Seed Classification from the Visible and Near-Infrared Spectrum
Countries to which the data refer | GERMANY | |
Countries to which the data refer | FRANCE | |
Countries to which the data refer | SERBIA | |
References to related material | https://max-kuk.github.io/bisid-5k/ | |
Type of the data | Dataset | |
Type of the data | Image | |
Total size of the dataset | 1070234765956 | |
Author | Kukushkin, Maksim | |
Author | Bogdan, Martin | |
Author | Goertz, Simon | |
Author | Callsen, Jan-Ole | |
Author | Oldenburg, Eric | |
Author | Enders, Matthias | |
Author | Schmid, Thomas | |
Upload date | 2025-04-10T08:11:53Z | |
Publication date | 2025-04-10T08:11:53Z | |
Data of data creation | 2025-03-15 | |
Publication date | 2025-04-10 | |
Abstract of the dataset | The success of deep learning in image classification has been largely underpinned by large-scale datasets, such as ImageNet, which have significantly advanced multi-class classification for RGB and grayscale images. However, datasets that capture spectral information beyond the visible spectrum remain scarce, despite their high potential, especially in agriculture, medicine, and remote sensing. To address this gap in the agricultural domain, we present BiSID-5k, a thoroughly curated bimodal seed image dataset comprising paired RGB and hyperspectral images for 10 plant species, making it one of the largest bimodal seed datasets available. We describe the methodology for data collection and preprocessing and benchmark several deep learning models on the dataset to evaluate their multi-class classification performance. By contributing a high-quality dataset, BiSID-5k offers a valuable resource for studying spectral, spatial, and morphological properties of seeds, opening new avenues for research and applications. | |
Public reference to this page | https://opara.zih.tu-dresden.de/handle/123456789/1410 | |
Public reference to this page | https://doi.org/10.25532/OPARA-810 | |
Publisher | Universität Leipzig | |
Licence | Attribution-NonCommercial-NoDerivatives 4.0 International | en |
URI of the licence text | http://creativecommons.org/licenses/by-nc-nd/4.0/ | |
Specification of the discipline(s) | 2::23::207 | |
Title of the dataset | BiSID-5k: A Bimodal Image Dataset for Seed Classification from the Visible and Near-Infrared Spectrum | |
Research instruments | Resonon Pika L 100121-220 | |
Research instruments | Sony IMX477R, 12.3 MP | |
Project abstract | In the project “KIRa - AI-supported platform for the classification and sorting of plant seeds: Evaluation of seed purity using rapeseed as a sample case”, state-of-the-art machine learning methods and a robotic sensor and sorting solution are being researched, combined as a platform and iteratively expanded as a ‘learning machine’. Through this platform, the KIRa sorter, we will automate and digitize the legally required purity testing in seed production as an important part of agriculture and develop it as a cooperative process between users and AI. This will significantly increase efficiency compared to manual testing and improve the quality and, in particular, the purity of seeds. The resulting KIRa sorter could thus represent a product with high potential for commercial exploitation in the future, whereby the individual components and their research and development will also yield innovative scientific results. | |
Public project website(s) | https://www.npz-innovation.de/projectKIRA.html | |
Project title | KIRa - AI-supported platform for classifying and sorting plant seeds: Evaluation of seed purity using rapeseed as a sample case |
Files
License bundle
- Name:
- license.txt
- Size:
- 4.66 KB
- Format:
- Item-specific license agreed to upon submission
- Description:
Collections
