Towards Closing the Loop in Robotic Pollination for Indoor Farming via Autonomous Microscopic Inspection

TL;DR

This work tackles the challenge of autonomous pollination in indoor farming by integrating a robotic manipulator with a vibrating pollination tool, an endoscope, and an autofocusing microscope to close the loop from global flower detection to micro-scale pollen verification. The approach combines RGB-D perception, 6-DoF pose estimation, visual servoing, and in situ pollen inspection to both pollinate and verify deposition, enabling real-time quality control. Key contributions include open-source hardware/software for the end effector, robust multi-scale perception and control pipelines, and extensive subsystem and full-pipeline validation on artificial and real strawberry flowers, including a demonstration of improved fruit yield via buzz pollination. The results demonstrate the feasibility of bringing a microscope to the sample and autonomously verifying pollination, with practical impact for indoor farming where traditional pollinators are unavailable.

Abstract

Effective pollination is a key challenge for indoor farming, since bees struggle to navigate without the sun. While a variety of robotic system solutions have been proposed, it remains difficult to autonomously check that a flower has been sufficiently pollinated to produce high-quality fruit, which is especially critical for self-pollinating crops such as strawberries. To this end, this work proposes a novel robotic system for indoor farming. The proposed hardware combines a 7-degree-of-freedom (DOF) manipulator arm with a custom end-effector, comprised of an endoscope camera, a 2-DOF microscope subsystem, and a custom vibrating pollination tool; this is paired with algorithms to detect and estimate the pose of strawberry flowers, navigate to each flower, pollinate using the tool, and inspect with the microscope. The key novelty is vibrating the flower from below while simultaneously inspecting with a microscope from above. Each subsystem is validated via extensive experiments.

Figures (7)

• Figure 1: Our proposed robotic manipulator system performs autonomous pollination and microscopic inspection of strawberry plants as follows. First, the global scope subsystem estimates flower locations with RGB-D camera. Then, the local scope subsystem uses a particular flower location as a reference to estimate the 6D flower pose and align the arm for contact. Next, the microscope subsystem uses visual servoing to facilitate contact with the flower, then alternates between inspection and buzz pollination with our custom, open-source pollination tool. Finally, once pollination is confirmed, the robot moves to the next flower.
• Figure 2: Overview of our proposed autonomous robotic pollination system.
• Figure 3: An overview of our custom end effector.
• Figure 4: Close-up view of our pollination tool before and during contact with a real strawberry flower. The cup shapes on the tines at the distal end of the tool help to grasp the flower.
• Figure 5: Visualization of autofocus operation, which stops once the focus score surpasses the threshold. Time points 1, 2, and 3 match to the microscope views at the top.