10.60507/FK2/DHTEH1 Halstead, Michael Michael Halstead https://orcid.org/0000-0001-7185-9304 University of Bonn Guclu, Esra Esra Guclu https://orcid.org/0000-0003-1106-5566 University of Bonn Farag, Mohamed Mohamed Farag https://orcid.org/0000-0003-4301-1140 University of Bonn Pallotta, Enrico Enrico Pallotta https://orcid.org/0009-0003-9276-0540 University of Bonn Hund, Christian Christian Hund University of Bonn Roscher, Ribana Ribana Roscher https://orcid.org/0000-0003-0094-6210 University of Bonn Bennewitz, Maren Maren Bennewitz https://orcid.org/0000-0003-4343-3028 University of Bonn Gall, Juergen Juergen Gall https://orcid.org/0000-0002-9447-3399 University of Bonn McCool, Chris Chris McCool https://orcid.org/0000-0002-0577-1299 CSIRO bonndata 2026 Agricultural Sciences Güclü, Esra Esra Güclü University of Bonn 2026-07-14 2026-07-17 2021-08-18/2021-09-10 Images; Numerical; Spreadsheet; Structured; Qualitative 10.48550/ARXIV.2607.14934 10.60507/FK2/DHTEH1/WIZCY9 10.60507/FK2/DHTEH1/LG9IR3 10.60507/FK2/DHTEH1/N5XPHK 500703195 5739 5419 application/gzip text/plain text/markdown 1.0 Creative Commons Attribution 4.0 International License. BUTom21 dataset is a fully hand-annotated still image dataset of tomatoes. It contains 123, 72, 98 images respectively in the training, validation, and evaluation subsets. The dataset contains RGB and noisy depth images captured across 2 capture days and two different RealSense D435i cameras. This creates a fully pixel-wise annotated dataset for tomato detection and segmentation in a real-world robotics scenario.