How kcsumesh25/cloud_masking works
This repository is a small tutorial collection for OmniCloudMask, not the library implementation: it contains three notebooks and a README. The documented product workflow is to prepare imagery with Red, Green, and NIR bands, run cloud and shadow inference, and obtain a four-class segmentation mask, optionally with confidence output. Notebook names identify Sentinel-2, WorldView, and DEA Sentinel-2 examples, but their contents were not available, so their internal steps and calls cannot be verified. Product capabilities below are documented in the README; implementation links are omitted where evidence is absent.
Main entry points: 01_s2_cloud_shadows_masking.ipynb Sentinel-2 example · 02_wv_cloud_shadows_masking.ipynb WorldView example · 03_s2_dea_cloud_shadows_masking.ipynb DEA Sentinel-2
Components 9
Tutorial notebooks
- Sentinel-2 example
01_s2_cloud_shadows_masking.ipynb - WorldView example
02_wv_cloud_shadows_masking.ipynb - DEA Sentinel-2
03_s2_dea_cloud_shadows_masking.ipynb
Masking product
- OmniCloudMask
- Four-class mask
- Confidence map
Other components
- Imagery analyst
- Satellite imagery
- Inference hardware
Connections 7
- Imagery analyst → Sentinel-2 example: opens example
- Imagery analyst → WorldView example: opens example
- Imagery analyst → DEA Sentinel-2: opens example
- Satellite imagery → OmniCloudMask: input bands
- OmniCloudMask → Four-class mask: produces mask
README.md - OmniCloudMask → Confidence map: optionally exports
README.md - Inference hardware → OmniCloudMask: supports execution
README.md