Developing an Integrated Solution for Apple Bud Thinning with Computer Vision and Buds Removal Mechanisms
Research Poster Engineering 2025 Graduate ExhibitionPresentation by Kittiphum Pawikhum
Exhibition Number 187
Abstract
Apple crop thinning is critical for optimizing fruit quality, yield consistency, and orchard productivity. Traditional manual bud thinning is labor-intensive, costly, and subject to variability. Recent advances in computer vision and robotics offer opportunities to automate this task efficiently and precisely. This study presents an integrated robotic system combining machine vision and robotics for automated apple bud thinning. The primary objectives were to: (1) develop a vision-based system to detect apple buds accurately; and (2) design a robotic mechanism capable of precisely removing the identified buds. A machine vision system was developed using a deep learning model (YOLO) for accurate bud detection from RGB images. Additionally, a 3D neural network trained on point cloud data was implemented to estimate optimal grasping poses for the robotic manipulator. The robotic removal mechanism was designed, prototyped, and integrated with the vision system, enabling precise, automated bud thinning. The integrated system was tested in laboratory and orchard environments to evaluate accuracy, efficiency, and robustness. Preliminary results demonstrated that the robotic system achieved high accuracy in bud detection and reliable removal in both controlled and real-world conditions. Performance metrics, including detection accuracy, thinning efficiency, and system reliability, were quantified, showing significant improvements compared to manual methods. Ongoing work focuses on optimizing system performance and addressing practical challenges for broader commercial adoption. This research advances precision agriculture by providing an automated, scalable solution for apple bud thinning, significantly enhancing orchard management and productivity through the effective integration of computer vision and robotics.
Importance
This research has the potential to address a critical need in apple production by automating the labor-intensive bud thinning process, traditionally performed manually. By integrating advanced computer vision and robotics, this technology could significantly reduce labor costs, enhance thinning accuracy, and support consistent fruit quality and yield. Automating this process may offer orchard managers reliable, scalable, and efficient tools, potentially reducing reliance on manual labor. This study aims to improve orchard productivity and economic outcomes and demonstrates the potential for broader implementation of robotic systems in precision agriculture, laying the groundwork for future automated crop management strategies.