Tao, Colleagues Describe Robotic Orchard Maintenance at Ag Conference

Yang Tao

Professor Yang Tao presented research on robotic orchard thinning at an automation workshop held as part of the Penn State University Fruit Research and Extension Center’s Grower Field Day this summer. Tao and his colleagues, Paul Heineman and Jude Liu of Penn State University, described a system in which a combination of computer vision, artificial intelligence, and a robotic arm could simulate the decision making process of a human performing thinning on fruit trees.

“Thinning” is a process in which a grower removes blossoms or small, immature fruit from a tree to reduce the ultimate number of mature fruits it will produce. While a large harvest may seem appealing, more is not always better–if every piece of fruit is allowed to grow and ripen, their overall quality and size will be reduced, and the crop is more likely to be attacked by destructive insects. Thinning encourages quality over quantity.

It is, however, a time-consuming task. Thinning machinery does exist, but it lacks the decision-making power of the grower’s mind–it is unable to asses the best young fruit or blossoms to leave in place.

Tao, whose research at the Bio-Imaging and Machine Vision Laboratory covers areas including food quality and safety, robotics and automation, worked with Heineman and Liu to propose an automated thinning system that simulates human decision-making and movements. A combination of computer, robotic arm and camera, mounted on a mobile platform, would move up and down a row a trees. The camera would take pictures of the trees, which would then be evaluated by the computer. When the end of a branch is detected, the system would measure back to its origin to determine its length, count the number of blossoms, use an equation based on the length of the branch and number of blossoms to determine how many blossoms should be removed, and send instructions to the robotic arm to grip the branch and do the job.

For More Information:

Visit the Bio-Imaging and Machine Vision Laboratory web site »

Published August 25, 2009