Unlike satellite images, which are typically acquired and processed in near-real-time, global land cover products have historically been produced on an annual basis, often with substantial lag times between image processing and dataset release. We developed a new automated approach for globally consistent, high resolution, near real-time (NRT) land use land cover (LULC) classification leveraging deep learning on 10 m Sentinel-2 imagery. We utilize a highly scalable cloud-based system to apply this approach and provide an open, continuous feed of LULC predictions in parallel with Sentinel-2 acquisitions. This first-of-its-kind NRT product, which we collectively refer to as Dynamic World, accommodates a variety of user needs ranging from extremely up-to-date LULC data to custom global composites representing user-specified date ranges. Furthermore, the continuous nature of the product’s outputs enables refinement, extension, and even redefinition of the LULC classification. In combination, these unique attributes enable unprecedented flexibility for a diverse community of users across a variety of disciplines.

No abstract

No abstract

This technical note describes the data sources and methodology underpinning a computer system for the automated generation of land use/land cover (LULC) maps of urban areas based on medium-resolution (10–30m/pixel) satellite imagery. The system and maps deploy the LULC taxonomy of the Atlas of Urban Expansion—2016 Edition: open, nonresidential, roads, and four types of residential space. We used supervised machine learning techniques to apply this taxonomy at scale. Distinguishing between recognizable, clearly defined types of land use within a built-up area, rather than merely delineating artificial land cover, enables a huge variety of potential applications for policy, planning, and research. We demonstrate the training and application of machine-learning-based algorithms to characterize LULC over a large spatial and temporal range in a way that avoids many of the onerous constraints and expenses of the traditional LULC mapping process: manual identification and classification of features. This document supersedes the previous technical note, Spatial Characterization of Urban Land Use through Machine Learning, and the methodology described here supersedes our previously reported techniques.

No abstract