An Integrated Actuation-Perception Framework for Robotic Leaf Retrieval: Detection, Localization, and Cutting

Robotic Assessment of a Crop’s Need for Watering: Automating a Time-Consuming Task to Support Sustainable Agriculture

“…We can perform this process autonomously by integrating four core functionalities: 1) end-effector design to cleanly cut and retain leaves, 2) onboard visual perception to detect and localize candidate leaves, 3) robot arm manipulation to enclose a candidate leaf with the end effector and cut it, and 4) waypoint navigation of the wheeled mobile robot based in the field to travel from one sampling location to another and back to the SWP analysis station. Our previous work [3] has focused on studying the third functionality, namely, co-optimizing perception and actuation for autonomous leaf retrieval in static cases. Here, we present research findings on the important components of the development and assessment of the end effector, camera selection, and assessment to enable visual leaf identification and pose estimation as well as overall system integration and field deployment and testing.…”