Oil Radish Growth Dataset
News
2019.08.26 – Codalab competition up again
Codalab, which is hosting the Oil Radish Growth Dataset competition, experienced a major crash a while ago. They have restored their system, but from older backups. This unfortunately means that we have had to recreate the competition, which means it now has a new competition ID.
The Oil Radish Growth Dataset competition can now be found here: https://competitions.codalab.org/competitions/20981
Data
The data includes pixel-wise annotated images of the sampling areas, field measurements of the sampling areas, daily weather data and unlabelled images from the full plots.
| Data | Description |
| Images (train) | Training images cropped to match the sampling areas in plot 1-3. |
| Labels (train) | Training labels associated with the training images. |
| Images (test) | Test images cropped to match the sampling areas in plot 4. The associated labels are not publicly available, but evaluation is possible through the competition site. |
| Unlabelled images | Non-cropped unlabelled images collected from plot 1-3. This data can be used for semi- or un-supervised training. |
| Field data | The field data includes fresh weight, dry weight, C-content and N-content for each fraction for each sampling area. The fractions are oil radish, barley, stubble and weed. |
| Weather data | The weather data includes daily measurements of air and soil temperature, precipatation, and global radiation from sowing date to the last day of image collection. |
The labelled data (full dataset excluding unlabelled images) as well as starting kit is also available through the online competition at CodaLab.
Challenges
The Oil Radish Growth Dataset is accompanied by two challenges: 1) The Semantic Segmentation challenge and 2) the Yield Estimation challenge. The challenges are detailed in the following sections.
To participate in the challenges, download the dataset, create a model for either or both of the challenges, and submit it to theonline competition.
Semantic Segmentation challenge
<Semantic segmentation challenge description>
Yield Estimation challenge
<Yield estimation challenge description>