Real-world agriculture datasets are scarce and hard to obtain, especially with a novel RGB-D sensors. This new data-set provides an important contribution to the research community, that will benefit from the viticulture dataset collected, which we believe can enhance future developments. Data included scans of grape clusters on real vine plants acquired in different illumination outdoor and vineyard conditions, as acquired by a RGB-D sensor mounted on a mobile robotic platform.
More details about the dataset in our upcoming publication here. Please kindly consider referencing to the publication if you find the data useful for your research:
Polina Kurtser, Ola Ringdahl, Nati Rotstein, Ron Berenstein, Yael Edan. Unsupervised learning for in-field grape cluster size assessment using a mobile robot and an RGB-D camera. IEEE Robotics and Automation Letters (2020). DOI: 10.1109/LRA.2020.2970654
Presented at ICRA 2020 (May 31-June 4th, Paris, France, remotely, presentation available here for those who attended ICRA)
The data is available for download: here and here
The data includes 10 video recordings made by an Intel Realsense camera in .bag format according to the protocols described in the paper in outdoor and vineyard conditions.
If you have additional questions about the dataset, write to us: polina.kurtser{at}oru{dot}se or natirot{at}post{dot}bgu{dot}ac{dot}il