Total Recall: Counting All Visible and Hidden Flowering, Underripe, and Mature Strawberries Over the Course of a Growing Season

US strawberry growers face a worsening labor shortage that is threating to become a crisis. This has resulted in grower operating expenses increasing faster than the market rate for strawberries andis eroding their already thin margins. Without solutions, the bulk of the industry is likely to leave the US, but L5 Automation is working to prevent that undesirable outcome. In this SBIR project, L5 will research key components of a future system that will produce much more accurate yield forecasts than is feasibile today - particularly in the 4-6 week window critical for negotiating purchase agreements with berry buyers. More accurate yield forecasts will allow growers to sell a higher percentage of their crop ahead of time at better prices and more precisely anticipate their labor needs. In turn, thishelps protect their margins and reduces waste.This project falls under USDA SBIR Topic Area 8.13 Plant Production and Protection, specifically improved harvesting of specialty crops and development of cyber-physical systems to support efficient agriculture.The autonomous robotic systems used and improved in this research project will regularly surveil strawberry fields to track plants over time andcount the flowers and strawberries present - key metricsneeded for an accurateyield forcasting system. Today, it would not make financial sense to build such a system based on the efforts human laborers and while systems using drones or rovers have been attempted, they are insufficiently accurate as they cannot count what is hidden under the foliage as they lack the ability to safely manipulate the plants. L5has alreadyproven that its robots have the ability to gently interact with the plants, moving leaves out of the way to expose the entire plant down to the bed surface. This capability will be key to collecting the accurate counts of all the flowers and strawberries present, ultimately supportint significantly more accurate yield forecasting.