Agricultural Robotics to Revolutionize Farming

Abstract

Innovations in terms of robotic manipulator control and field robots in digital agriculture have advanced considerably in the last decade, with the aim of reducing costs and increasing efficiencies. The availability of compact imaging sensors, such as digital cameras that can perceive depth information, besides the flexibility of open-source image processing software that can be trained for different applications has played significant roles in accelerating this sector. The presented survey summarizes some of the recent advances in redundant manipulators that are controlled using image-based visual servoing (IBVS) for automating various farming tasks including (1) pruning, thinning, and trimming, (2) harvesting, and (3) inspection and target spraying. The study also covers advances in field robots and their availability in the European market. The reviewed works suggest that developing optimal tree shapes and planting techniques is necessary to improve the performance of visual servo control and automate farming operations with robots. In addition, selecting the right imaging sensors, employing graphics processing units, and training the computer vision algorithms with more fruit and plant datasets have been highlighted as the three main elements for improving the functionality of IBVS in manipulator control for agricultural applications.