Field validation of an autonomous electrostatic drone pollination system for Japanese pear

Artificial pollination is essential in Japanese pear ( Pyrus pyrifolia ) cultivation due to widespread self-incompatibility; however, conventional manual pollination is labor-intensive and increasingly unsustainable under current labor shortages. This study presents the development and field validation of a fully autonomous drone-based pollination system integrating AI-based flower cluster detection, RTK-GNSS navigation, and electrostatic pollen spraying (−12 kV). Field trials were conducted in a joint V-shaped orchard canopy, comparing electrostatic drone pollination, nonelectrostatic drone pollination, manual feather-brush pollination, and natural pollination. The electrostatic drone treatment achieved a fruit set rate of 62.1%, comparable to manual pollination (61.9%), while reducing operation time to 9.3 s m⁻¹ and decreasing pollen consumption relative to conventional methods. In contrast, the nonelectrostatic drone treatment showed lower fruit quality performance, indicating the contribution of electrostatic charging to targeted pollen deposition. Distance-dependent analysis demonstrated improved spatial stability of good fruit production under electrostatic conditions. A two-dimensional finite element analysis further clarified that grounding the drone body enhanced electric field concentration toward flower clusters, supporting efficient pollen attraction. These results demonstrate that the proposed autonomous electrostatic drone system can achieve pollination performance comparable to conventional manual methods while improving labor and resource efficiency. The study provides practical design insights for integrating AI-guided navigation and electrostatic spraying in orchard-based smart pollination systems.