Growth Analysis of Wheat Using Machine Vision: Opportunities and Challenges

Area of Science:

• Plant science
• Agricultural technology
• Genetics

Background:

• Continuous crop growth analysis is crucial for yield potential and stress tolerance assessment.
• Traditional methods require numerous replicates and destructive measurements, limiting scalability.
• Machine vision presents a promising non-destructive alternative for high-throughput crop phenotyping.

Purpose of the Study:

• To evaluate machine vision for inferring plant growth parameters in spring wheat.
• To assess the correlation between machine vision-derived traits and traditional destructive measurements.
• To determine the utility of machine vision for high-throughput genetic analysis.

Main Methods:

• Utilized machine vision and RGB imaging to capture side-projected area (SPA) as a high-throughput trait.
• Performed destructive measurements for biomass (BIO), leaf dry weight (LDW), and leaf area (LA) at multiple time points.
• Analyzed growth parameters and relative growth rates using both machine vision and destructive data.

Main Results:

• Significant effects of time and genotype were observed on all measured traits (BIO, LDW, LA, SPA).
• Side-projected area (SPA) strongly predicted leaf area, leaf dry weight, and biomass.
• Relative growth rate estimated via SPA robustly predicted rates derived from biomass and leaf dry weight.

Conclusions:

• Machine vision, specifically RGB imaging for SPA, is a reliable and efficient non-destructive method for crop growth analysis.
• This approach enables the assessment of large numbers of genotypes for genetic mapping, generating continuous growth curves with reduced replication.
• SPA serves as a valuable surrogate trait for key plant growth parameters in high-throughput phenotyping.