A Visual Recognition and Path Planning Method for Intelligent Fruit-Picking Robots

With the rapid development of economy and increasing improvement agricultural production level, people’s demand for fruits is also year by year. China largest fruit consumption country in world. According to relevant statistics reported China, end 2019, total amount various sold had reached about 270 million tons, with apples accounting 48% global output pears 69% national output. However, China’s picking still dominated manual process, which takes a lot manpower time complete, therefore resulting low efficiency. Some farmers are unable complete short time, large number be listed, huge losses. To solve this problem, paper focuses on visual recognition path planning intelligent fruit-picking robot. Using robot best way at present. This establishes positioning system based stereo vision, used identify locate planted orchard area. The coordinate error target point less than 10 mm, has high accuracy. Then, planned feedback algorithm biological stimulation neural network. Our empirical evaluations suggest that proposed walks planting park shape “zigzag” realizes full-coverage after 6 turns. results show efficiency method.