The imagery data comes from a variety of sources. The bulk of the data was collected in-field by CIMMYT and CIMMYT partners in Ethiopia and Tanzania. The remainder of the data are sourced from public images found on Google Images.
The aim of this challenge is to build a machine learning model to accurately classify the wheat in the images as: healthy, stem rust, or leaf rust.
Some images may contain both stem and leaf rust, there is always one type of rust that is more dominant than the other, i.e. you will not find images where both appear equally. The goal is to classify the image according to the type of wheat rust that appears most prominently in the image. The values for each classification can be between 0 and 1, inclusive, and should represent the probability that the wheat in the image belongs in each category.
There are 876 images to train your model on and 610 images in the test set by which your solution will be evaluated on Zindi.
Files available for download:
You can read more about wheat rust here. Note that Stem rust occurs primarily on stems but can also be found on leaves, sheaths, glumes, awns, and even seed. Leaf rust is generally found on leaves but may also infect glumes and awns.