Frontiers in plant science · 2026
YOLOv8-PnP fusion architecture for non-contact robotic pollination: a 6D pose estimation approach for autonomous greenhouse operations.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Singh Rajmeet, Kaur Manveen, Gadade Appaso M, Hussain Irfan
Аннотация
BACKGROUND: The declining availability of natural pollinators and limitations of contact-based robotic pollination methods including flower damage, pathogen transmission, and reduced operational efficiency necessitate innovative solutions for protected horticulture.
GAP: Existing robotic pollinators achieve limited success rates (∼ 66%) primarily due to inaccurate 6D flower pose estimation, while current airflow-based systems lack precise positioning capabilities.
CONTRIBUTION: This study presents a novel YOLOv8-PnP hybrid framework integrating real-time object detection with 6 degree of freedom pose estimation for precision airflow based pollination. The system employs a custom-designed Air Pollenmatic end-effector integrated with a Hello Robot Stretch platform through ROS-based visual servoing control.
RESULTS: Validation on 2,100 annotated greenhouse images demonstrated 95.8% precision, 94.6% recall, and 97.7% mAP@0.5 at 28.5 FPS (11.1 ms inference). Field trials achieved 92.5% pollination attempt rate and 85.6% success rate, yielding 79.2% overall efficacy-an 8.3 percentage point improvement over contact-based methods.
IMPACT: This contactless approach eliminates mechanical flower damage, reduces disease transmission risk, and advances the feasibility of fully autonomous greenhouse pollination systems.
GAP: Existing robotic pollinators achieve limited success rates (∼ 66%) primarily due to inaccurate 6D flower pose estimation, while current airflow-based systems lack precise positioning capabilities.
CONTRIBUTION: This study presents a novel YOLOv8-PnP hybrid framework integrating real-time object detection with 6 degree of freedom pose estimation for precision airflow based pollination. The system employs a custom-designed Air Pollenmatic end-effector integrated with a Hello Robot Stretch platform through ROS-based visual servoing control.
RESULTS: Validation on 2,100 annotated greenhouse images demonstrated 95.8% precision, 94.6% recall, and 97.7% mAP@0.5 at 28.5 FPS (11.1 ms inference). Field trials achieved 92.5% pollination attempt rate and 85.6% success rate, yielding 79.2% overall efficacy-an 8.3 percentage point improvement over contact-based methods.
IMPACT: This contactless approach eliminates mechanical flower damage, reduces disease transmission risk, and advances the feasibility of fully autonomous greenhouse pollination systems.