Geometric Feature Characterization of Apple Trees from 3D LiDAR Point Cloud Data.

Area of Science:

• Agricultural Engineering
• Remote Sensing
• Horticulture

Background:

• Effective orchard management relies on accurate geometric feature characterization of fruit trees.
• LiDAR (light detection and ranging) technology offers rapid and precise evaluation of these features.
• Quantifying tree height, canopy volume, and spacing is crucial for optimizing orchard practices.

Purpose of the Study:

• To quantify key geometric features of apple trees in an orchard using a 3D LiDAR sensor.
• To assess the accuracy of LiDAR-derived measurements for tree height, canopy volume, and spacing.
• To evaluate the potential of LiDAR technology for enhancing orchard management strategies.

Main Methods:

• Collected 3D point cloud data from an apple orchard using a LiDAR sensor.
• Processed LiDAR data using commercial software and Python, including outlier removal, downsampling, denoising, segmentation, and ground point removal.
• Validated sensor-estimated geometric features against manually measured values for six apple trees.

Main Results:

• High accuracy was observed for tree height (r²=0.98, CCC=0.96) and canopy volume (r²=0.97) estimations.
• Tree spacing (r²=0.92) and row spacing (r²=0.94) were also accurately measured.
• Minor differences between sensor estimates and measurements indicate efficiency, with potential for refinement.

Conclusions:

• 3D LiDAR technology provides accurate geometric feature characterization for apple orchards.
• The findings support the use of LiDAR for precise measurements aiding orchard management.
• Further validation with larger, diverse datasets is recommended to enhance accuracy and generalizability.