Written by Ranjan Sapkota and Manoj Karkee, Center for Precision and Automated Agricultural Systems (CPAAS), Washington State University (WSU), 99350, USA; January 31, 2024.
The decline in natural pollinators, particularly bees, poses a significant challenge for apple production. Bees play a crucial role in food production, including apples, as they are responsible for one out of every three bites of food that we eat around the world [1]. However, factors such as habitat loss, pesticide use, and climate change have led to a decrease in bee populations globally. This decline threatens the sustainability production of high quality apples in different parts of the world. In response, scientists and agriculturalists, including our team at the WSU Center for Precision and Automated Agricultural Systems (CPAAS), are exploring alternative pollination methods to ensure consistent and efficient pollination in different fruit crops including apple orchards.
Robotic Pollination System
A robotic pollination system was designed at WSU CPAAS as an advanced alternative to traditional pollination methods. The robotic pollinator comprises a low cost (Intel RealSense) camera, and a robotic manipulator arm (Universal Robots, Odense, Denmark) installed on a robotic ground vehicle called a Warthog (Clearpaths Robotics, Ontario, Canada). The system also includes an end-effector or tool for precisely delivering pollens to target flowers. Photos collected with the Intel RealSense camera from the apple orchard are processed using two different Artificial Intelligence (AI) models called Mask-RCNN and YOLOv8 to identify and locate target flowers accurately. The manipulator arm, which has six joints to move it around, provides the necessary range of motion to position the spray tool precisely. The spray tool consisting of an electrostatic spray nozzle applies a specially formulated pollen solution to the flowers.
Figure 1: Showcasing our advanced robotic pollination system in a commercial orchard. It features an RGB-D (Red, Green, Blue – Depth) camera for collecting canopy images, a UR5e manipulator arm mounted on a Warthog ground robot by Clearpath Robotics, and an efficient electrostatic sprayer system for targeted pollination.
Field Study and Results
In 2022, our study used a Mask R-CNN model to detect and locate flowers in photos captured with the Intel RealSense camera. The locations of flowers were then sent to the robot, which moved the spray tool to target flowers resulting in a 56% success rate in pollinating flower clusters. In 2023, we improved the image analysis model by using YOLOv8 and focusing exclusively on detecting and locating king flowers to align with the goal of ‘one fruit per flower cluster.’ This advancement of the robotic system led to substantial improvements in flower detection (91% of king flowers detected), and the pollination success rate jumped to 84%. Additionally, the cycle time, the duration from taking photos of the canopies to pollination, was reduced from 6.5 seconds to 4.8 seconds per flower cluster. These advancements highlighted the system’s enhanced efficiency and precision, suggesting its potential to for practical adoption in the future.
Conclusions, Improvements, and Future Potential
The recent study at WSU CPAAS on robotic pollination of apples showcases substantial progress and a potential for an alternative pollination approach. In the last a few years, the technology evolved substantially as evidenced by improved detection accuracy and pollination success rates. These advancements lay the groundwork for further development and potential widespread adoption to commercial apples and other fruit crop farming. Wider commercial adoption of this technology has the potential to transform the way apples and other fruit crops are managed. It is also expected, when successfully adopted, that robotic pollination of targeted flowers will avoid the need for flower thinning and minimize the need for fruitlet thinning in orchards.
Contact
Center for Precision and Automated Agricultural Systems, Washington State University, 99350, USA
manoj.karkee@wsu.edu
509-786-9208
Funding and acknowledgements
The research and development of robotic pollination system presented in this article was funded by USDA’s Specialty Crop Multi-state grant program.
Additional Information
- Pollinator Partnership. (n.d.). Pollinators need you. You need pollinators: Why are pollinators important?.
Fruit Matters articles may only be republished with prior author permission © Washington State University. Reprint articles with permission must include: Originally published by Washington State Tree Fruit Extension Fruit Matters at treefruit.wsu.edu and a link to the original article.
Use pesticides with care. Apply them only to plants, animals, or sites listed on the labels. When mixing and applying pesticides, follow all label precautions to protect yourself and others around you. It is a violation of the law to disregard label directions. If pesticides are spilled on skin or clothing, remove clothing and wash skin thoroughly. Store pesticides in their original containers and keep them out of the reach of children, pets, and livestock.
YOU ARE REQUIRED BY LAW TO FOLLOW THE LABEL. It is a legal document. Always read the label before using any pesticide. You, the grower, are responsible for safe pesticide use. Trade (brand) names are provided for your reference only. No discrimination is intended, and other pesticides with the same active ingredient may be suitable. No endorsement is implied.
