"A million numbers in a spreadsheet is not a product," Orchard Robotics founder and CEO Charlie Wu told the Ruggedize ag robotics conference at Reservoir Farms in Salinas, California. "Data is only useful if a grower has a mechanism of taking an action on that data easily. Your average grower does not want to be, nor should they be, a data analyst."

Orchard Robotics makes an AI-powered camera about the size of a large tissue box, weighing around five pounds, that mounts on any tractor or farm vehicle and captures millions of images as it passes through orchards and vineyards. It measures fruit counts, size and colour, growth rates, canopy health, water stress and disease, and estimates yield. Images are processed on board on an NVIDIA Jetson computer because many farms lack the connectivity to upload terabytes of imagery; the results go into the company's FruitScope platform, where growers can look at individual trees.

The simplest use is fruit count, Wu said. In tree fruit, growers prune and thin to adjust the number of fruit per tree, and within a block half the trees may carry double the fruit of the other half — so knowing where labour is needed, and where it is not, is the first saving. "The biggest ROI is through reduction on labor and inputs," he told AgFunderNews. Yield forecasts at 95% accuracy or better then feed supply-chain planning: how many pickers to hire, how many bins to buy, how much storage to book and what the sales team can promise.

Farms with precision sprayers and spreaders can take the platform's insights directly; the rest receive instructions through a mobile app that tells workers where to go and what to do. Growers scan anywhere from five to 25 times a season, with about ten the average. Wu said the company does not promise higher yields when selling: "We pitch them on the labor savings, the input savings, and supply chain efficiencies" — and a grower who captures all of those in the first year can see a three to ten times return.

The business runs on a per-acre, per-year subscription that includes the software, unlimited scans and the hardware, so growers get each new camera version free rather than being left with an obsolete unit. Some customers put the camera on their own tractors during mowing or spraying passes; others use a full-service option in which Orchard's technicians scan a field and deliver the data the next day.

Orchard started in apples and now works in wine and table grapes, cherries, blueberries, almonds, pistachios and citrus, and is moving into pomegranates, coffee and strawberries. Wu said the AI models learn across crops and improve with each new farm. The company has hundreds of systems in the field scanning tens of thousands of acres, aims for hundreds of thousands, and while most customers are in the US it is seeing "some pretty tremendous pull internationally".

Source: AgFunderNews