In this notebook I use the SentinelHub statistical API to gather NDVI data in a 3 year timespan for two agricultural fields in Flanders region.

Download the source .TIFF file from this link

• Copy this file to data/sentinel2/source

Download the source geodata for agricultural areas in Flanders from this link

• Extract this file into data/agriculture_geodata

Create a virtual environment with Python 3.10

Install the necessary libraries with

pip install -r requirements.txt

• Run the Jupyter notebook sentinel-stat-ndvi.ipynb in your favorite IDE.

External scripts, classes and functions are located in the scripts folder

Outputs from UTM conversion will be saved in the data/processed folder

Sentinel requests for specific fields will be saved in the data/sentinel2/fields folder

Outputs from the violin-charts will be saved in the data/export folder

1. Obtaining the labels
2. Conversion to UTM31
3. Selecting two fields
4. Obtaining statistical data from Sentinel2
5. Cleaning the data
   • Cloud cover
   • Cloud probability
   • Snow cover
6. Result
7. Violin charts

We obtain the following chart of the NDVI of the two fields for a 3 years time period.

This notebook accompanies my article on Medium

This notebook was written by Frank Trioen, April 2023.