The Vision-Based Target Recognition, Localization, and Control for Harvesting Robots: A Review

In recent years, the elderly population has increased, leading to a labor shortage and the increasing cost of training experienced labor. Owing to the continuous optimization of machine vision, multi-sensor technologies, control methods, and end-effector structures, harvesting robots have experienced rapid development. However, most harvesting robots still require intelligent solutions, and the lack of integration with artificial intelligence limits them to small-scale applications without mass production. This paper reviews key technologies for vision-based sensing and control of harvesting robots, focusing on potential applications of vision for target recognition and localization in complex agricultural environments, describing improved solutions for different target detection and localization algorithms, and comparing their detection results. The challenges and future trends of applying these key vision sensing and control techniques in harvesting robots are also described and discussed in this review.