Fruit Harvesting

Robotic fruit harvesting aims to automate fruit picking, addressing labor shortages and improving efficiency in agriculture. Current research heavily emphasizes computer vision, employing deep learning models (like YOLOv5 and various 3D object detection networks) for fruit detection, pose estimation, and peduncle localization, often integrated with advanced robotic arm designs (including novel hybrid architectures) and sophisticated control algorithms. These advancements, including the development of large, publicly available datasets like MetaFruit, are improving the accuracy and speed of robotic harvesting, with promising results demonstrated across various fruit types in both controlled and field settings.