Vision-Based Robotic Grasping Method for Nest Frame Based on Gen6D and AnyGrasp

Beekeeping plays a significant role in agricultural production, while the traditional farming method makes beekeepers not only have heavy physical labor, but also easy to be attacked by bees. Nest frame grasping is the core task of honey harvesting robots, for which a vision-based robotic grasping method combining Gen6D with AnyGrasp is proposed. After that, an evaluation and decision system based on the nest frame mechanics model is added to select the most suitable grasping pose. Finally, vision program tests and grasping experimental based on a mobile base 6-DOF manipulator are carried out. The method succeeds in 35 gripping experiments under 4 different working conditions, with acceptable response time, specifically, 10.632 s on average. The study contributes to the application of honey harvesting robots, saving response time and improving grasping accuracy significantly.