Sentinel-2 COG analysis - NDVI calculator

This repo will help you install a Python based CLI program, that uses Element84 API to acquire Sentinel-2 L2A images from AWS Open Data S3 buckets, in your area of interest and time range, to calculate the NDVI values in an area of interest.

The output JSON response contains average, median, min and max NDVI to allow better understanding of the Areas of Interest.

The provided Jupyter Notebook sen2_ndvi.ipynb allows the user to visualize the outputs of the Python scripts, on a Folium basemap.

Note : Currently the Area of Interest is hard-coded to be the GeoJSON acquired from this link. The link right now provides a Multipolygon GeoJSON and the script performs NDVI calculation and KMeans classification on all the Polygons in a Multipolygon.

Setting up the Environment

This script was developed using Python > = 3.6, in Ubuntu 20.04.

Note : osgeo gdal must be setup, and the script gdalinstall.sh helps installing gdal in a Debian OS such as Ubuntu. It is run as below -

bash gdalinstall.sh

The following packages must be available in the Python installation -

• requests
• numpy
• GDAL
• rasterio
• PIL
• pyproj
• shapely

The above packages can be installed using the requirements.txt file using :

pip install requirements.txt

How to run the Script, and what to expect

The main.py script runs the code to extract NDVI.

A JSON response containing NDVI statistics, links to color image, and the name of the Polygon that is processed is saved as analysis_response.json.

The main.py script requires --start_date and --end_date as arguments.

To run the script, use the terminal to navigate to the cloned repo, and checkout the --help option -

python3 main.py --help

Output -

usage: main.py [-h] [--start_date START_DATE] [--end_date END_DATE]

optional arguments:
  -h, --help            show this help message and exit
  --start_date START_DATE
                        Provide a start date as string in YYYY-MM-DD format
  --end_date END_DATE   Provide an end date as string in YYYY-MM-DD format

Example:

python main.py --start_date 2021-11-07 --end_date 2021-10-05

The above command runs the scripts and gives logs in the terminal as the script proceeds, and finishes with the below output.

NDVI analysis complete :

Response : [{'polygon_id': 'poly_1', 'image_date': '2021-10-28', 'true_color_image': 'https://sentinel-cogs.s3.us-west-2.amazonaws.com/sentinel-s2-l2a-cogs/33/U/UU/2021/10/S2B_33UUU_20211028_0_L2A/TCI.tif', 'average_ndvi': 0.6183, 'median_ndvi': 0.6898, 'min_ndvi': -1.0, 'max_ndvi': 0.9972}, {'polygon_id': 'poly_2', 'image_date': '2021-10-28', 'true_color_image': 'https://sentinel-cogs.s3.us-west-2.amazonaws.com/sentinel-s2-l2a-cogs/33/U/UU/2021/10/S2B_33UUU_20211028_0_L2A/TCI.tif', 'average_ndvi': 0.6656, 'median_ndvi': 0.6692, 'min_ndvi': -0.0679, 'max_ndvi': 0.9967}]

---

Testing

Codes have been added to test the STAC response filtering with the help of Mock Data.

The tests can be run with -

python3 tests/test.py