Dataset fOr herding and predatoR detectIon with the use of robotS (DORIS)

Dataset with images and annotations (semantic masks for YOLO and other architectures) of sheep, wolves, persons and depth images in order to carry out experiments of robots used for herding sheep and detect potential predators as wolves. It is divided into two different URLs since it is greater than 50GB. This dataset contains images of persons, wolves, sheep and the depth images that can be used to simulated environments and they have the depth estimated from the images using Depth Anything. We gratefully acknowledge the financial support of Grant TED2021-132356B-I00 funded by MCIN/AEI/10.13039/501100011033 and by the "European Union NextGenerationEU/PRTR".

Since the original dataset exceeds 50 GB (https://doi.org/10.57967/hf/2059), there are three datasets associated:

a dataset that contains YOLO annotations for these images.(this dataset)

a second dataset contains images of Sheep, (10.5281/zenodo.11313800)

a third dataset contains Person, Wolf classes, and depth maps generated using Depth Anything, which can be utilized for simulating external environments or outdoor navigation. ( 10.5281/zenodo.11313966)