Land Cover Classification

Land cover and land use information underpins natural resource management, sustainable development and environmental impact mitigation. EOMAP's Land Cover Classification derives thematic land cover and land use maps from satellite imagery without stepping foot on the ground, from rapid identification of single classes to complex habitat mapping of vast areas.

Each project starts by agreeing the level of detail and a classification scheme with the client: an existing nomenclature such as CORINE Land Cover, the FAO Land Cover Classification System or ESA WorldCover, or a scheme tailored to the site. EOMAP then selects imagery to match. Sentinel-2 and Landsat 8 give 10–30 m products. Commercial very-high-resolution data gives finer detail at additional data cost: the product page quotes 2–5 m from providers such as Maxar and Airbus, and EOMAP's sensor table lists Maxar WorldView, Pléiades Neo, Planet SuperDove and SkySat, and SPOT 6 at 1–6 m. Images are atmospherically corrected and spectrally filtered, then classified object by object on spectra, shape, texture and change over time, optionally sped up with AI, followed by calibration and validation. EOMAP reports that clients typically validate its classifications at 85–90% accuracy.

Results are delivered as vector shapefiles with a PDF map and metadata; grids and other formats are available on request. Land cover can be merged with EOMAP's seafloor classification into one seamless coastal dataset. In Florida, as part of a wetland habitat mapping product launched with Navico in 2017, EOMAP has classified almost 20 wetland macrophyte species on lakes such as Lake Istokpoga on behalf of the Florida Fish and Wildlife Conservation Commission.