Multi-Resolution Deep Learning for Land Use Applications

Global production of perishable food crops is becoming increasingly vulnerable to supply chain interruptions given climatic changes in key growing areas, with associated pressures such as reduced water stores, increasing incidents of new diseases and pests, and disruptive events such as wildfires and floods. Detailed knowledge of crop supply at any given time and an ability to forecast the anticipated supply into the near future are key to increasing the resiliency, efficiency and sustainability of perishable food crop production into the future. This Phase II Extension proposes to refine capabilities developed during Phase II to determine current day supply of several important vegetable and berry crops and to develop a capability to forecast their future supply. We use novel techniques to classify the crops, which are grown in many relatively small fields in close proximity and with relatively short and staggered growth periods - attributes which historically have prevented their reliable detection and tracking with satellite imagery. We apply an innovative approach to generating larger sets of labeled data given the typical dearth of labels from on-the-ground information, which promises to make our capability highly scalable and efficient. Applying machine learning and computer vision techniques to time series of satellite imagery, we expect to be able to forecast harvest events, on a field by field basis, several weeks in advance and with an error of about a week. Aggregating field areas for each crop and applying yield-per-area estimates for each crop type, we expect these performance metrics will provide sufficiently accurate yield forecasts to generate commercial demand by sectors of the fresh produce industry and to demonstrate capabilities with potential applications to global food security.

Related follow-on opportunities for NASA program infusion include integration with the TOPS-SIMs irrigation management program at the Ecological Forecasting Lab at NASA Ames, and NASA Goddard’s Harvest Consortium led by the University of Maryland to enhance the use of satellite data in decision making related to food security and agriculture, and the Surface Biology and Geology (SBG) Decadal Designated Observable Study.

Related commercialization opportunities include monitoring and forecasting for industrial agriculture, particularly for fresh vegetable crops, improved cropland classification for USDA’s Cropland Data Layer, and food waste and sustainability applications addressing prioritized actions of the EPA, USDA and FDA.

This is early/mid-stage (TRL 6) — the most realistic path in is NASA SBIR/STTR, which funds small businesses and research institutions to develop technology aligned with NASA's needs (equity-free, phased funding). Check whether a current SBIR/STTR solicitation topic overlaps with this project's technology area, or contact the project directly (above) to ask.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.