This repository provides curation of the DIST-S1 model.
See how this data is used to train a vision transformer to detect surface disturbances in dist-s1-model.
This repository:
- localizes the OPERA RTC-S1 data from the ASF DAAC
- Finds spatial patches within a fixed Sentinel-1 burst that are used to construct co-registered time-series small enough to be analyzed by the vision transformer
- Serializes the npz files for each data sample
- Validates and visualizes the sample dataset.
- Loads the data files into torch.
Go through the notebooks.