Row selection in remote sensing from four-row plots of maize and sorghum based on repeatability and predictive...

Area of Science:

• Agronomy
• Plant Breeding
• Remote Sensing

Background:

• Remote sensing offers rapid trait assessment for plant breeding, enhancing genetic gain.
• Traits are typically extracted from row segments, allowing for quantitative analysis of subsets within plots.

Purpose of the Study:

• To evaluate the impact of row selection and plot trimming on remote sensing trait analysis.
• To determine optimal methodologies for field trials using multi-sensor data.

Main Methods:

• Utilized RGB, LiDAR, and VNIR hyperspectral data from uncrewed aerial vehicle flights over sorghum and maize trials.
• Analyzed traits based on all rows, inner rows, outer rows, and individual rows.
• Assessed the effect of 40 cm plot end trimming on data repeatability and yield prediction.

Main Results:

• Plot trimming did not significantly alter outcomes compared to non-trimmed plots.
• Row selection significantly impacted results, with more row segments generally increasing repeatability.
• Excluding outer rows enhanced the predictive modeling of end-season yield.

Conclusions:

• Row selection is a critical factor in remote sensing-based phenotyping for plant breeding.
• Excluding outer rows can improve yield prediction models, aligning with established agronomic practices.
• These findings support the integration of refined remote sensing methodologies into breeding programs.