UAV-based individual plant detection and geometric parameter extraction in vineyards

Area of Science:

• Agricultural Engineering
• Horticulture
• Remote Sensing

Background:

• Precise vineyard management requires accurate characterization of macroscopic vineyard parameters.
• Traditional manual measurements are often time-consuming and not representative of the entire vineyard row.

Purpose of the Study:

• To present a novel approach for detecting trunk positions and extracting macroscopic vineyard characteristics using point cloud data.
• To enable informed management decisions regarding pesticide application, defoliation, and biomass assessment for optimal sugar content.

Main Methods:

• Utilized point cloud data for trunk detection and macroscopic vineyard characteristic extraction.
• Employed a method based solely on geometric features, compatible with various training systems and 3D sensors.
• Conducted extensive experiments on grapevine rows trained in two different systems.

Main Results:

• The proposed method accurately and efficiently extracts plant height, canopy width, and canopy volume.
• The approach provides more comprehensive canopy characteristics compared to traditional manual measurements.
• Demonstrated effectiveness and robustness across different grapevine training systems.

Conclusions:

• The novel point cloud data approach offers a valuable tool for precise vineyard management and breeding.
• This method enhances yield monitoring, grape quality optimization, and strategic interventions for increased vineyard productivity and sustainability.