MIT Technology Review: Daily satellite imagery is useful for tracking a variety of information about Earth's surface, including rainfall, crop growth, vegetation patterns, and flood and wildfire damage. However, a clear image of the surface is often hard to obtain because of cloud cover. Descartes Labs, a startup founded to commercialize software developed for satellite and aerial imagery at Los Alamos National Laboratory, has developed a new algorithm that combines daily satellite images and automatically edits out the clouds. By comparing color, IR, and UV satellite imagery with annotated maps, the software can identify agriculture details, water features, and types of forest. From that information, the software has been trained to predict future corn crops based on the color and appearance of the plants in a field. According to Steven Brumby, Descartes Labs cofounder and chief technology officer, the software's predictions, which were based on 11 years of satellite imagery, were more accurate than the predictions made by the US Department of Agriculture, which were based on data collected directly from farmers.
