Unsupervised Deep Learning for Irrigation Detection in Satellite Imagery

Our mission is to build the best model we can that will be used by the California Department of Food and Agriculture to sustainably manage the most critical resource known to humankind — freshwater.

Our team is focused on using multi-spectral satellite imagery from the Sentinel-2 satellites to detect irrigation behaviors in California. The problem of detecting irrigation resources from satellite images is largely unstudied, but it is a compelling problem to solve because of the potential applications including studying the effects of climate on agricultural water sources across the world, providing unbiased measurements of water resource usage, and detecting water theft and illegal agriculture.