Achieving Robotically Peeled Lettuce

Abstract

Robotic technologies are being increasingly applied to agriculture, in particular to harvesting. Some types of produce such as iceberg lettuce require additional processing after harvesting in order to satisfy the needs of the end-user or customer. Lettuce must have its outer leaves removed, a task that is currently performed manually. The leaf removal task represents a challenging vision and manipulation problem: the lettuce is in a random pose on a flat surface, from which the outermost leaves must be removed quickly and without causing damage. This letter presents a novel vision pipeline and suction removal system that enables robotic lettuce leaf removal. A suction nozzle and control procedure are used for the removal itself, relying on the orientation estimation and stem detection provided by the vision pipeline. To the best of the author's knowledge, this is the first robotic lettuce leaf peeling system.