2010 Volume 72 Issue 2 Pages 152-159
Experiments on a cherry tomato harvesting robot for a mobile cultivation facility were performed. In this report, the harvesting end effector, image recognition in order to avoid obstacles, and harvesting experiments are discussed. To harvest the fruit along with its calyx, the harvesting end effector was equipped with a semicircular component and a press band. The end effector grasped the peduncle between these components and detached the fruit at the joint. Five paths were programmed in order to avoid collisions. Obstructions in each harvesting path were recognized by processing 3-D image, and the path with the fewest obstructions was selected. Of the total number of fruits, 81% were successfully harvested, and 98% of the harvested fruits had an intact calyx.