MN Impact: Engineers and Breeders Team-Up to Develop New Technologies for Apple Orchards
Collaborative research will lead to new tools to help specialty crop growers make better decisions regarding labor and sales. (Reported as the percent accuracy achieved by the new tool).
Specialty crop growers rely on manual labor for fruit picking, inspecting, data collecting and other labor-intensive tasks. The less structured environments inherent with specialty crops systems have made automation more difficult but the increasing cost and decreased availability of seasonal workers makes such advances increasingly important.
What has been done
Ibrahim Volkan Isler and U of M apple experts James Luby and Cindy Tong have teamed up to develop technology that will focus on two tasks: counting apples in an orchard and measuring their diameter. Counting apples accurately can be difficult, even for humans, due to variations in skin color, inconsistent light conditions and obstructions of fruits by leaves, branches and other fruit.
The team developed deep learning and image geometry based methods to accurately analyze the visual imagery. This new approach resulted in 95.56 to 97.83 percent yield accuracies. Other technologies, including a robotic arm to help with picking, are in development.
A start-up company called Farm Vision Technologies was launched in 2017 with an aim of integrating this and other technology into farms to help farmers monitor count, size and health of their plants. Ultimately, this technology will help farmers make decisions regarding labor and sales far in advance, directly improving their revenue and reducing their costs.