The tech world is witnessing rapid growth in Geographic Information System (GIS) applications powered by satellite data. The data community is increasingly relying upon these space-bound instruments to solve problems related to environmental monitoring, hazard detection, asset management and more.
Our series on Satellite data applications will help you get started with understanding data, performing analysis and building interesting applications through demos.
Let’s start with what satellite data can represent. The most obvious data are the optical images of the Earth. Other than that, satellites also capture:
- SAR Images (Sentinel-1)
- Atmospheric Data (Sentinel-5P)
- Digital Surface Model / Digital Elevation Model
For starters, we shall discuss all about optical images in this blog. Optical images are captured as RGBI images. Sentinel program provides free RGBI images (Sentinel-2) to play with. The data is provided as Raster.
A Raster is a file format that stores a bunch of metadata in addition to the image channels. Crucially, it connects the exact boundaries of the images to a location on earth. It also stores a coordinate reference system (CRS). There are a number of ways locations of spherical earth are projected on a 2D, and a coordinate reference system helps not just standardize the methods, but also allows conversion between them.
Commercial raster is often accompanied by a shapefile. A shapefile is very similar to CSV data frame. Importantly, it has a special column called geometry. The geometry column can contain spatial features like points, lines, polygons, etc. The features are, again, geo-referenced. The GeoPandas library of Python helps us deal with shapefiles, in a manner like Pandas dealing with CSVs.
It’s more efficient to run mathematical operations like union, intersection, masking on a shapefile than raster, as the former is the order of magnitude smaller in size.
QGIS is an open-source tool that helps us map and visualize rasters and shapefiles. It also comes with a ton of plugins like maps, base maps, Google and Bing satellite images, street view, etc.
A view of the QGIS application
In the next article, we will see how to get data and load it up for analysis. Stay tuned.