Sky Kit

Easy to use out-of-the-box satellite imagery package for Python aficionado.

Adding fun to satellites coolness 😎

Warning: This is still a work in progress. Please do not use in production, yet.

Seriously, what is it? 🧐

Sky kit is an ensemble of methods to ease satellite imagery processing, from and to developers and data scientists. Working with any satellite imagery provider, should be as easy as:

from skykit.providers import <ProviderName>

Install 🚀

As with any python package, pip is your friend:

pip install skykit

Usage 🎡

Quick Start

The following shows details of a tile from Sentinel-2, that includes Paris (France), from 01/03/2019 to 08/03/2019:

from datetime import date
from skykit.providers import Sentinel2

sat = Sentinel2(username="xxx", password="yyy")

tiles = sat.query(
    coordinates=(2.349014, 48.864716),     # Paris (FR) coordinates
    dates=(date(2019,3,1), date(2019,3,8)) # From march 1, 2019 to march 8, 2019
    )

len(tiles) # returns 6

tile = tiles[0]         # Pick any tile you want, or filter more
print(tile['summary'])  # returns 'Date: 2019-03-02T10:50:29.024Z, Instrument: MSI, Mode: , Satellite: Sentinel-2, Size: 730.27 MB'

In the previous example, the tiles variable will include meta information of every tile returned from Sentinel-2 query.

Features

Connecting to satellite images providers

Actually, Sky kit support the following ticked providers (others coming):

• Sentinel 1
• Sentinel 2
• Landsat 7
• Landsat 8
• Modis
• Cbers-4
• Aqua

All image providers are grouped into skykit.providers module. Invoking a provider (like Sentinel-2) can be done like this:

from skykit.providers import Sentinel2

We then need to initialize the connection with the provider. eg with Sentinel-2:

sat = Sentinel2(username="xxx", password="yyy")

Nota: Sentinel 1 & 2 require having credentials to connect to Copernicus Open Access Hub. Once created, insert your credentials in the code above.

Todo 🔭

• Write a better documentation. Look for readthedocs.io
• Add providers
• Create a reusable abstraction of Tile, to create a uniform use of tiles, no matter which provider
• Ease working with tiles meta data (usually about the scene and satellite)
• Allow querying using a polygon
• Allow querying using an address
• Use friendly dates (like text: "26/03/2019"), of use from="..." and to="..." instead of date=(...)
• Allow working on specific bands. Something like:

image = tile.get('b04') # or tile.b04() ???

• More tests <-- ⚠ URGENT
• Create a CI/CD pipeline to publish tags into PyPi

This list is not exhaustive.

Contribution 💪

Please suggest contributions using Github's Pull Requests.

License 🤮

Read the LICENSE.txt file.

Thanks 🙏