Approaches for Improving Fruit Detection and Gripping Mechanisms in Orchard Robotic Fruit Harvesting

Harvesting robots provide a promising solution to overcome the challenges of manual harvesting in fruit orchards, which is usually the most time-consuming and laborious part of the orchard fruit production process. Despite the technological advancements in robotic harvesting, real-world applications are hindered by field-level limitations. This chapter first analyzes the major challenges in field-level harvesting robotic applications concerning the detection power of the robotic vision system and robotic arm manipulation and then proposes corresponding state-of-the-art developments to overcome these challenges. The key challenges encountered include (1) occlusion caused by leaves, fruit clusters, branches, and other plant parts, which reduces both the detection ability of the robotic vision system and the operational capacity of the robotic gripper, and (2) day–night light variation (uneven lighting), wind, and reflections or shadows that limit the detection ability of the vision system. There are two types of approaches to robotic fruit harvesting. On the one hand, developments in horticulture have compelled the use of simpler harvesting robots due to the simpler fruit architecture in contemporary fruit orchards than in conventional orchards. On the other hand, recent developments in electronics and computational capacity can make the task of robotic operations easier, especially in orchard fruit picking. Recently developed vision systems are reliable and accurate in localizing target fruits. Within this context, the application of simple linear and rotational manipulations in orchard fruit harvesting robots has become increasingly important as an alternative to existing complex harvesting robots.