Predicting Zea mays Flowering Time, Yield, and Kernel Dimensions by Analyzing Aerial Images

Area of Science:

• Agricultural Science
• Remote Sensing
• Plant Phenotyping

Background:

• Traditional crop phenotyping methods can be labor-intensive and less efficient.
• Image analysis offers a promising alternative for objective and high-throughput crop assessment.

Purpose of the Study:

• To evaluate the efficacy of unmanned aerial vehicle (UAV)-based vegetation index imaging for maize phenotyping.
• To correlate vegetation index data with key crop traits like flowering time, yield, and kernel dimensions.

Main Methods:

• Utilized a recreational-grade UAV with a modified camera to capture Blue Normalized Difference Vegetation Index (BNDVI) images weekly over three seasons.
• Applied Principal Components Analysis (PCA) to analyze BNDVI histogram changes and Partial Least Squares Regression (PLSR) for predictive modeling.

Main Results:

• BNDVI values and histogram shapes effectively tracked crop growth stages, from canopy closure to senescence.
• PCA components correlated with flowering time (PC2, PC3) and yield (PC2).
• Mid-season BNDVI positively correlated with yield, particularly with tall, thin kernels, and PLSR accurately predicted flowering time and yield.

Conclusions:

• UAV-based BNDVI imaging provides a cost-effective and efficient platform for maize phenotyping.
• This approach can reliably predict critical crop traits, aiding in breeding and management decisions.