Sentinel2Py

Sentinel2Py is a Python package for downloading, stacking, and visualizing Sentinel-2 satellite imagery. It integrates STAC search via Microsoft Planetary Computer and provides utilities to generate RGB composites and spectral indices such as NDVI and NDWI.

---

Features

• Search for Sentinel-2 L2A tiles by bounding box, date range, and cloud cover
• Download bands using presets (RGB, RGBNIR, NDVI, NDWI, etc.)
• Stack bands into single GeoTIFFs at native or specified resolution
• Plot RGB, NDVI, and NDWI with percentile stretching, gamma correction, histogram equalization, and optional normalization
• Download multiple tiles in one call

---

Installation

git clone https://github.com/capizziemanuele/sentinel2py.git
cd sentinel2py
pip install -r requirements.txt

---

Initialize Manager and Plotter

from sentinel2py.downloader.manager import Sentinel2Manager
from sentinel2py.plot.plotter import SentinelPlotter

manager = Sentinel2Manager(out_dir="./data")
plotter = SentinelPlotter()

---

Download & Stack Bands

downloaded, stacked = manager.download_bands(tile, preset="RGBNIR", stack=True)
stack_path = list(stacked.values())[0]  # Get stacked file path

---

Plot RGB and Indices

# RGB True Color
plotter.plot_rgb(
    stack_path,
    bands=(3,2,1),  # B04,B03,B02
    downsample=4,
    stretch=True,
    gamma=1,
    equalize=False)

# NDVI
plotter.plot_ndvi(
    stack_path,
    bands=(4,3),    # NIR, RED
    downsample=4,
    stretch=True,
    normalize=False,
    cmap="RdYlGn")

# NDWI
plotter.plot_ndwi(
    stack_path,
    bands=(2,4),    # GREEN, NIR
    downsample=4,
    stretch=False,
    normalize=False,
    cmap="Blues",
    equalize=True)

---

Download Multiple Tiles

all_downloaded, all_stacked = manager.download_multiple_tiles(
    tiles,
    preset="RGBNIR",
    stack=True,
    target_res="highest",
    overwrite=False)

---

Optional: Cloud Cover Table

import pandas as pd

df = pd.DataFrame([{
    "Tile": t.id.split("_")[-2],
    "Date": t.properties.get("datetime").split("T")[0],
    "Cloud (%)": t.properties.get("eo:cloud_cover", "N/A")
} for t in tiles])

print(df)

---

Notes & License

---

Notes

• NDVI and NDWI values can remain in the natural range [-1,1] or be normalized to [0,1]
• BAND_PRESETS are customizable; you can define new combinations of bands
• Downsampling (downsample) is useful for plotting large tiles without using too much memory.