Study on Cartesian-Type Strawberry-Harvesting Robot

摘要

Harvesting strawberries requires high labor costs and long working hours. In order to reduce the labor burden in agriculture, robotic automation is very much desired. Especially, there is a strong demand for the automated harvesting and grading of ripe and soft strawberries, as such takes tend to take long working hours and requires special attention not to damage the fruits. A Cartesian-type strawberry-harvesting robot was developed for in-row hilltop culture. The robot consists of two color cameras (wide-view and detail cameras), a harvesting hand, a fiber-optic sensor and a mobile control unit. The robot can detect the fruit, evaluate its ripeness and pick strawberries by cutting its peduncles. Experimental performance tests showed that the robot could accurately detect the fruits (93.6% efficiency) and harvest samples with more than 50% ripeness level. Some difficulty was observed in accurately detecting the fruit peduncles (70.8% efficiency), Execution time for successfully harvesting a single fruit was about 16.6 s. The result of the testing indicated the potential of harvesting robot with a harvesting hand as well as automatic harvesting algorithm function.