October 29, 2010   

A group of researchers at Carnegie Mellon University is looking to revolutionize one of the most basic elements of civilization.
Work is under way on projects aimed at bringing robotics to agriculture. One project in particular, which would introduce robotic smarts into the $3 billion domestic grape industry, is just kicking off.
The work isn’t designed to bring robo-pickers into the vineyard. Instead, it’s aimed at using high-tech imaging to accurately estimate crop yield and canopy.
“This is fundamental to our existence. It makes sense to inject science,” said Debadeepta Dey, a doctoral student at CMU and one of the project’s collaborators. “We grow crops on such a large scale, and the labor doesn’t scale.”
Grape estimates are carried out in a labor intensive and inefficient manner that also can be destructive to the vines, the researchers say. Leaves on a row of vines are counted by hand, then multiplied across acres of a vineyard, or leaves are plucked off, weighed and multiplied. Either way, the results may be less specific than needed.
Collaborators — which include researchers at the CMU Robotics Institute, Intel Lab Pittsburgh, Cornell University and the industry group National Wine and Grape Initiative — started a pilot project over the summer that would tackle this element of grape growing.
Yield estimates are important because they lay the foundation for harvest plans, said Lily Mummert, research scientist at Intel Labs Pittsburgh. The team is looking at multiple ways to solve the problem and intends to apply for Department of Agriculture funds early next year to ramp up the project.

Link to article