EarthDefine Land Cover

EarthDefine Land Cover is an on-demand product that maps high-resolution land cover for any area in the contiguous United States. Our proprietary AI delivers over 95% accurate 60cm land cover, typically within days of an order, and can produce historical snapshots by classifying older aerial imagery from our archive.

We deploy robust, purpose-built models that extract each ground-cover class from high-resolution colour-infrared imagery, at continental scale and with consistent, repeatable results. Because we train an independent model for each class, a single patch of ground can be captured under every type it truly belongs to. This way, overlapping features like trees over impervious surfaces or trees over buildings are resolved correctly rather than forced into one label. The same precision captures fine detail (individual trees, narrow strips of bare ground, small water bodies) that coarser datasets average away.

• measuring current and potential tree canopy for Urban Tree Canopy (UTC) assessments
• understanding stormwater runoff quantity and quality from the distribution of impervious and other land cover
• improving line-of-sight radio-wave propagation modelling for wireless planning via a derivative clutter classification
• conducting environmental impact assessments for construction projects
• monitoring land-cover change related to oil and gas development
• finding least-cost pathways for utility corridors and highways
• predicting erosion risk by combining land cover with terrain data
• wildfire risk modelling and post-fire ecological impact assessment
• monitoring urban growth
• improving population surface maps with dasymetric mapping for disaster risk analysis
• modelling carbon storage and change as a function of land-cover type