SB20
SB20 is a sugar beet dataset that was captured at a field on campus Klein Altendorf of the University of Bonn. The data was captured by mounting an Intel RealSense D435i sensor with a nadir view of the ground on a Sagarobot. It contains RGB-D images of sugar beet and eight different categories of weeds covering a range of growth stages, natural world illumination conditions, and challenging occlusions. The dataset provides multi-class annotations.
Download the dataset here. Also, cite the following papers if you used this dataset:
@article{DBLP:journals/corr/abs-2109-11936,
author = {Alireza Ahmadi and Michael Halstead and Chris McCool},
title = {Towards Autonomous Crop-Agnostic Visual Navigation in Arable Fields},
journal = {CoRR},
year = {2021},
url = {https://arxiv.org/abs/2109.11936},
}
@article{DBLP:journals/corr/abs-2106-10118,
author = {Alireza Ahmadi and Michael Halstead and Chris McCool},
title = {Virtual Temporal Samples for Recurrent Neural Networks: applied to semantic segmentation in agriculture},
journal = {CoRR},
year = {2021},
url = {https://arxiv.org/abs/2106.10118},
}