gdalcubes is an R package and C++ library aiming at making processing collections of satellite imagery easier, faster, and more interactive.

It provides functions to create and process four-dimensional regular raster data cubes from irregular image collections, hiding complexities in the data such as different map projections and spatial overlaps of images, or different spatial resolutions of spectral bands.


  • Create data cubes from local files or cloud storage
  • Process data cubes by applying arithmetic operations on pixels (e.g. to compute vegetation indexes), computing summary statistics over space and time, filtering by space, time, and bands, and applying user-defined R functions on data cubes
  • Extract time series or pixels from spatial polygons and points
  • Export data cubes as netCDF or GeoTIFF file(s)
  • Create training data and apply predictions for machine learning models
  • Process data cubes in parallel while reading only smaller (memory-friendly) chunks of a cube at a time

Among others, gdalcubes has been successfully used to process data from the Sentinel-2, Sentinel-5P, Landsat, PlanetScope, MODIS, and Global Precipitation Measurement Earth observation satellites / missions.