For river centerline extraction and sediment bar identification from remote sensing images based on Google Earth Engine (GEE) Python API
- Water and sand identification.
- Automatic river identification from water mask.
- One-pixel wide river centerline extraction.
The algorithm is friendly to river planform and centerline extraction over expansive regions from images collected from multiple dates.
- Highly automated.
- Free cloud space for high-speed computation provided by GEE.
- Performed well on different imagery collections (PlanetScope and Landsat in the example).
- Require fewer spectral bands (RGB and near-infrared only).
|
|
- Sign up for Google Earth Engine
- Python >= 3.7
- NumPy
- Scipy
- earthengine-api
- geemap
- Jupyter Notebook It is recommended to install Miniconda, and create a virtual environment for GEE and other packages:
conda create -n geeCenterline -c conda-forge python=3 numpy scipy jupyterlab nb_conda_kernels earthengine-api geemap
conda activate geeCenterline
You can also install these packages by pip:
python -m pip install numpy scipy jupyterlab earthengine-api geemap
| Study area | Setellite | Date | ID |
|---|---|---|---|
| Little Tallahatchie River | PlanetScope | 10/7/2021 | 20211007_155117_27_245a |
| 20211007_155119_58_245a | |||
| 20211007_155121_88_245a | |||
| 20211007_155124_19_245a | |||
| 20211007_155126_49_245a | |||
| 20211007_155128_79_245a | |||
| 20211007_155131_10_245a | |||
| 20211007_155133_40_245a | |||
| 6/23/2022 | 20220623_155055_14_2465 | ||
| Landsat 8 | 6/25/2022 | LC08_L2SP_023036_20220625_20220706_02_T1 | |
| LC08_L2SP_023037_20220625_20220706_02_T1 | |||
| Big Sunflower River | PlanetScope | 8/13/2022 | 20220813_163511_16_2426 |
| 20220813_163513_42_2426 | |||
| 20220813_163515_68_2426 | |||
| 20220813_163517_93_2426 | |||
| 20220813_163520_19_2426 |
Can be obtained from Google Earth Engine following the steps in the example notebook.