Quantitative Analysis of Cotton Canopy Size in Field Conditions Using a Consumer-Grade RGB-D Camera

Area of Science:

• Agricultural Science
• Plant Phenotyping
• Computational Biology

Background:

• Plant canopy structure significantly influences crop yield and stress tolerance.
• Manual canopy size assessment is labor-intensive, imprecise, and limited to 1D measurements.
• High-throughput phenotyping systems offer rapid, quantitative field-based plant data acquisition.

Purpose of the Study:

• To develop and validate a 3D imaging approach for quantitative analysis of cotton canopy development in field settings.
• To assess the utility of multi-dimensional canopy traits for predicting cotton fiber yield.
• To identify optimal measurement times and traits for yield prediction in breeding programs.

Main Methods:

• Utilized GPhenoVision system for acquiring color and depth images of cotton plots.
• Developed a data processing pipeline: point cloud reconstruction, vegetation segmentation (using excess-green filter), and trait extraction.
• Extracted static morphological traits (height, width, area, volume) and growth rates; performed linear regressions with fiber yield.

Main Results:

• Fiber yield correlated significantly with static traits (R² = 0.35-0.71) post-canopy development and early growth rates (R² = 0.29-0.52).
• Multi-dimensional traits (projected area, volume) and the multivariate cumulative height profile outperformed univariate traits.
• The 3D imaging approach demonstrated effectiveness in quantifying cotton canopy development and its relation to yield.

Conclusions:

• The developed 3D imaging approach provides a robust method for quantitative cotton canopy analysis.
• Multi-dimensional and multivariate canopy traits are superior predictors of fiber yield compared to traditional univariate measures.
• This technology can aid in identifying quantitative trait loci (QTLs) and enhancing cotton breeding programs for improved yield prediction.