Application of Machine Vision on the Development of a Field Picking Robot for Passion Fruit

• Main Authors: Ping-Hung Chen, 陳秉鴻
• Other Authors: Feng Ou-Yang
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/15721847010547527177

碩士 === 國立宜蘭大學 === 生物機電工程學系碩士班 === 97 === The harvesting time of the passion fruits is from June to December. After June some passion fruits begin to get ripen and their surface color turns to brown from green. The mature fruit will fall on the ground from the shelf every afternoon. The farmers have to bend over to pick up the fruits from the ground and collecting in the buckets. The current harvesting method is quite time-consuming and cost-consuming. Long hours of bending over also induce waist injuries of farmers easily. This research has already developed a field mobile robot for picking up passion fruits. The robot is composed of the vision system and the mechanical control system. The vision system is able to recognize the fruit on the ground, and their locations will be calculated and passed to the mechanical control system to pick them up. The robot will move forward to next area automatically for next picking-up functions . The result of fruits recognition shows that the vision system can successfully recognize fruits and their numbers when fruits are not connected. When fruits are connected, the vision system can only recognize fruits but not the right number. The vision system could also recognize successfully under the following situations: light spots on fruits caused by the sunlight, shadow on fruits caused by the shelf and machine, weeds and stones on the ground. The experiment of robot arm picking function shows that 80% of fruits can be picked up successfully. The in-door experiment of the mobile robot shows that successful picking rate of the robot arm could be reach about 83% without any damage of the surface of the passion fruits.