Estimation of Characteristic Parameters of Grape Clusters Based on Point Cloud Data.

Area of Science:

• Agricultural Science
• Computer Vision
• Data Science

Background:

• Quantifying grapevine phenotypic parameters is vital for crop trait analysis.
• Individual variations in grape bunches complicate accurate parameter measurement.
• Existing methods struggle with the complexity of grape bunch geometry.

Purpose of the Study:

• To develop and validate a novel method for estimating grape feature parameters using point cloud data.
• To improve the accuracy of grapevine phenotypic parameter measurement.
• To provide a robust basis for the quantitative analysis of grape traits.

Main Methods:

• Grape point cloud segmentation using filtering and region growing algorithms.
• Complete grape point cloud model registration via an improved iterative closest point (ICP) algorithm.
• Phenotypic size characteristic estimation and surface reconstruction using the Poisson algorithm.

Main Results:

• The proposed method accurately estimates grape phenotypic parameters from point cloud data.
• Poisson surface reconstruction achieved a coefficient of determination (R²) of 0.9915.
• This R² value is significantly higher than the 0.7609 obtained with the alpha-shape algorithm.

Conclusions:

• The developed point cloud-based method offers superior accuracy for grapevine trait quantification.
• The Poisson reconstruction algorithm provides precise surface modeling for grape bunches.
• This approach establishes a strong foundation for detailed and accurate grape trait analysis.