batchsat

This project aims to build a software that can download Sentinel-2 satellite images in batch for data analysis.

how to install

with conda (we used miniconda3)

1. download the project
2. create a conda environment and run the following command: conda install rasterio=1.3.3 geopandas=0.12.1 -c conda-forge
3. then, run the command: pip install sentinelsat==1.1.1 xmltodict==0.13.0

how to use

create a csv file with the columns:

polygon_id;polygon_wkt;dt_from;dt_to
<unique_polygon_id_0>;<polygon_wkt_0>;<YYYY-mm-dd>;<YYYY-mm-dd>
.
.
.
<unique_polygon_id_1>;<polygon_wkt_1>;<YYYY-mm-dd>;<YYYY-mm-dd>

then, run python products.py --csv-path <path_to_csv_file>. there are other arguments, see with python products.py -h you will see that a products-polygons.csv was created.

use it as input to the next script, running: python rasters.py --csv-path <path_to_output_of_products.py>. if everything goes ok, it will be created a folder named rasters and subfolders named with the poygons unique id. inside these subfolders you will see the downloaded raster files ready to be used.

example

Example of a True Color Image 10m resolution time series from a random crop near Brasília, from 2022-06-03 to 2022-10-26.

project goals

• build a knowledge on how to collect, prepare, process and work with raster files and raster files time series
• learn digital and image processing techniques applied to satellite images
• provide a script to concurrently download images from Sentinel-2 twin satellites using computer with relatively good resources
• provide a container/microservice to scale the download satellite images in a k8s cluster
• help programmers and researchers to easily get satellite image data to make the world a better and safer place