Quick and easy indices assessment: exploiting the Google Earth Engine platform to detect peri-urban land cover changes

Vegetation indices, water indices, brightness and form indices, to name only a few, are long time classics for land cover use and land cover change detection. Prior to cloud computing, the standard workflow started from scenes selection by time, cloud filtering, image registration and, finally, indices evaluation. Working on local workstations, no matter how performant they are, other than being time-consuming, is often critical in terms of both computational load and mass storage requirements. Imagery fine-tuning still requires the possession of the physical files but cloud services can speed up to unprecedented levels most of the standard machinery of indices assessment. The Google Earth Engine platform allows quick and seamless access to the standard satellite imagery without downloading the actual scenes, thus providing the means to build time series of indices counting hundreds of records in almost no time. Furthermore, the Earth Engine platform supplies on-board raster algebra, so that it is not even necessary to download the indices for further calculations. The peri-urban landscape is characterised by land cover changes, often detectable through indices differences. The spatial scale needed by this kind of environment could benefit from the resolution of the current state of the art publicly available satellites, mainly the Sentinel-2 MSI and the Landsat-8 OLI sensors. At the price of some coarse-graining, older Landsat imagery is also available. We show that with a few lines of code users can highlight the putative land changes, creating a sketch land cover differential map.