Wheat grain width: a clue for re-exploring visual indicators of grain weight

Area of Science:

• Agricultural Science
• Plant Physiology
• Image Analysis

Background:

• Mean grain weight (MGW) is crucial in wheat breeding and physiology.
• Current image-based MGW estimation relies on limited indices, primarily projected area (Area).
• Automated image processing has advanced grain analysis but not MGW prediction precision.

Purpose of the Study:

• To explore novel image-derived indices for enhancing MGW estimation precision.
• To investigate the potential of grain width in improving MGW predictions.
• To develop and validate new predictive models for MGW.

Main Methods:

• Processed an image archive of over 72,000 wheat grains from a 2-year field experiment.
• Evaluated over 30 image-derived indices, including grain size and shape parameters.
• Developed and validated ten simple linear models for MGW prediction.

Main Results:

• Grain width indices (Minor & MinFeret) and Area/Perimeter showed higher correlation with MGW than Area.
• The most precise MGW predictions utilized Area×Circularity, Perimeter×Circularity, and Area/Perimeter.
• New models demonstrated higher precision than current Area-based models, with stable efficiency across environmental conditions.

Conclusions:

• Novel image-derived indices, particularly those incorporating grain width, significantly improve MGW prediction accuracy.
• Developed linear models offer enhanced precision for image-based MGW estimations.
• These advancements can facilitate more efficient wheat breeding and physiological assessments.