This repository contains my hands-on implementations of Google Earth Engine (Python API) workflows for remote sensing and geospatial analysis.
The notebooks demonstrate practical applications of cloud-based geospatial processing using satellite imagery and open geospatial datasets.
- Google Earth Engine (Python API)
- Geemap
- Landsat 8
- Sentinel-1
- Sentinel-2
- Image Collections
- NDVI & EVI
- Surface Water Mapping
- Agricultural Mapping
- Drought Monitoring
- Supervised Classification (Random Forest)
- Unsupervised Classification (K-Means)
- Accuracy Assessment
- Data Visualization
google-earth-engine-python-tutorials/
├── notebooks/
├── images/
├── data/
├── assets/
├── requirements.txt
└── README.md
Introduction to Google Earth Engine using the Python API. Learn how to authenticate, visualize satellite imagery, compute vegetation indices, and export raster datasets.
Topics Covered
- Google Earth Engine authentication and initialization
- Interactive mapping with Geemap
- SRTM Digital Elevation Model (DEM)
- Area of Interest (AOI)
- Landsat 8 imagery
- NDVI calculation
- EVI calculation
- Exporting GeoTIFFs
Notebook
notebooks/01_GEE_Basics_Python.ipynb
Outputs
Learn how to work with Landsat 8 ImageCollections by filtering satellite imagery based on location, acquisition date, and cloud cover, and create composite images for analysis.
Topics Covered
- Landsat 8 ImageCollections
- Spatial filtering (AOI)
- Date filtering
- Cloud cover filtering
- Image metadata
- Median Composite
- Mosaic
- Median vs. Mosaic comparison
Notebook
notebooks/02_Image_Collections_and_Cloud_Filtering.ipynb
Outputs
Notebook: 03_Cloud_Masking_&_Image_Preprocessing.ipynb
This notebook focuses on preprocessing Landsat 8 Collection 2 Level-2 imagery in Google Earth Engine before performing further remote-sensing analysis.
- Loading Landsat 8 Collection 2 Level-2 imagery
- Filtering images by date and area of interest (AOI)
- Filtering scenes using
CLOUD_COVER - Understanding the
QA_PIXELquality-assurance band - Detecting cloud and cloud-shadow pixels using bitwise operations
- Creating a cloud-masking function
- Applying the cloud mask to an
ImageCollection - Applying Landsat Collection 2 scale factors and offsets
- Creating median composites from cloud-masked imagery
- Comparing imagery before and after cloud masking
- Calculating NDVI using preprocessed imagery
Cloud filtering and cloud masking serve different purposes.
- Cloud filtering removes entire satellite scenes with high overall cloud cover.
- Cloud masking identifies cloudy pixels within the retained scenes and masks those pixels from further analysis.
The Landsat QA_PIXEL band is used to identify unwanted pixels such as clouds and cloud shadows.
Landsat Collection 2 Level-2 data are stored as scaled integer values. Scale factors and offsets are applied to convert the stored values into physically meaningful values.
| Product | Scale Factor | Offset |
|---|---|---|
Surface Reflectance (SR_B*) |
0.0000275 |
-0.2 |
Surface Temperature (ST_B*) |
0.00341802 |
149.0 |
General conversion:
Physical Value = DN × Scale Factor + Offset
This notebook demonstrates why satellite imagery should be preprocessed before spectral-index calculation and other remote-sensing analyses. Cloud filtering reduces the number of highly cloudy scenes, while pixel-level cloud masking removes contaminated observations from the remaining imagery.
- Python
- Google Earth Engine
- Geemap
- NumPy
- Pandas
- Matplotlib
- GeoPandas
- Rasterio
🚧 This repository is currently under development.
New notebooks will be added progressively, covering each topic in detail with explanations, code, and visual outputs.
This project is licensed under the MIT License.