IO Monitor Land Cover

Most publicly available global land-cover maps are refreshed only every one to two years, which leaves planners, governments and risk teams without a current view of how land is changing. IO Monitor Land Cover from Impact Observatory produces custom land use and land cover (LULC) maps on demand for any user-defined area of interest and any seasonal or annual date range, using deep-learning models that classify satellite imagery pixel by pixel.

The product comes in two editions. 10m Land Cover classifies Copernicus Sentinel-2 imagery into 15 classes, such as active and inactive cropland, dense and sparse trees and high- and low-density built area, with an archive from 2017 to the present. Each order includes the requested-period map, a baseline map of the same dates one year earlier, a change map and CSV metrics, delivered within hours. 3m Land Cover, built in partnership with Planet, combines daily PlanetScope imagery with Sentinel-2 to map 17 classes, adding buildings, roads, other built areas and mixed vegetation, from 2020 onward.

Maps are delivered as Cloud Optimized GeoTIFF and PNG files with pre-calculated metrics, by download or email, or transferred to Amazon AWS, Microsoft Azure or ArcGIS Online. The 10m edition uses the same model as the company's open Maps for Good data, for which it reports 85% accuracy. Published pricing is $2 per km² for 10m and $7.50 per km² for 3m, with tiered pricing for larger areas. Impact Observatory cites uses in urban planning, food security, supply-chain sourcing, water resources, disaster response and defence; Planet reports it delivered a map to the U.S. government within 24 hours ahead of Hurricane Idalia.