2010 Volume 72 Issue 5 Pages 479-486
We constructed a stationary robotic strawberry harvester consisting of an end-effector and manipulator with seven degrees of freedom (Yamamoto et al., 2009) and an artificial visual system for target fruit identification (Yamamoto et al., 2010). The stationary robot was tested in an experimental harvesting system in which the robot was combined with a movable bench unit. We examined the performance in 50 beds, each of which contained 10 strawberry plants. The success rate of harvesting mature fruit was 67.1%, the success rate of detecting fruit position was 89.0%, the rate of correct judgment of fruit coloration was 83.4%, and the success rate of picking detected fruit was 90.3%. The rate of damaged harvested fruit was 12.5% and the rate of picking the other fruits simultaneously was 13.9%. The average processing time per bed was 126.7 s when the average number of mature fruit in the bed was 2.3.
