README.md

Decloud

Decloud enables the training of various deep nets to remove clouds in optical images.

Representative illustrations:

Examples of de-clouded images using the single date SAR/Optical U-Net model.

Quickstart: Run a pre-trained model

Some pre-trained models are available. You can find more info on how to use them here

Advanced usage: Train you own models

1. Prepare the data: convert Sentinel-1 and Sentinel-2 images in the right format (see the documentation).
2. Create some Acquisition Layouts (.json files) describing how the images are acquired, ROIs for training and validation sites, and generate some TFRecord files containing the samples.
3. Train the network of your choice. The network keys for input/output must match the keys of the previously generated TFRecord files.
4. Perform the inference on real world images.

More info here.

Contact

You can contact remi cresson (Remi Cresson at INRAE )