Question: Interpretation of outputs in crop_segmentation.ipynb notebook about farmvibes-ai HOT 2 CLOSED

Hi, @jardabezdek. Thank you for raising this question.

In that example notebook, our goal is to train a crop land segmentation model. Given a set of Sentinel-2 images taken throughout a year, we want to estimate where the regions with crops are.

The maps that you see in the last cell show the prediction of the network (top image) and the ground-truth crop land map (bottom image). These are probability heatmaps, i.e., high values (yellow color) mean that region/pixel was assigned as crop land by the model, whereas low values (blue color) represent regions without crop areas.

We will improve on the description and labels of both maps for the next release.

from farmvibes-ai.

This was implemented in #14, specifically in https://github.com/microsoft/farmvibes-ai/blob/main/notebooks/crop_segmentation/04_inference.ipynb. I'm closing this issue.

@jardabezdek, please feel free to reopen if you think this could be further improved.

from farmvibes-ai.