A vision-based robotic system for precision pollination of apples
🤖 AI Summary
Global apple production faces mounting challenges due to declining natural pollinators and the low efficiency, poor robustness, and climate sensitivity of manual pollination. To address these issues, this work proposes a vision-guided robotic pollination system capable of single-flower-level precise manipulation—the first to achieve closed-loop autonomous control based on real-time floral organ recognition and 3D localization. The system integrates high-resolution multispectral imaging, an enhanced YOLOv8 model for organ detection, point cloud registration for spatial mapping, and coordinated control of a lightweight robotic manipulator to enable accurate, targeted pollen delivery to individual flowers. Field trials in commercial orchards demonstrate a pollination accuracy of 92.7%, an average operation time of less than 8 seconds per flower, and a 31% increase in fruit set compared to manual assisted pollination. This study establishes a novel paradigm for high-precision, autonomous agricultural robotics in complex, unstructured outdoor environments.
Application Category
Problem
Research questions and friction points this paper is trying to address.
Develops a robotic system for precision apple pollination.
Addresses decline in natural pollinators due to environmental factors.
Evaluates robotic pollination effectiveness in Honeycrisp and Fuji orchards.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Machine vision identifies flower clusters.
Manipulator guides sprayer for pollen application.
Robotic system achieves comparable fruit quality.
🔎 Similar Papers
No similar papers found.
