Stress Distribution Analysis on Hyperspectral Corn Leaf Images for Improved Phenotyping Quality

Area of Science:

• Agricultural Science
• Plant Biology
• Image Processing

Background:

• High-throughput imaging is crucial for agricultural plant phenotyping.
• Current methods average spectral data, failing to capture nutrient/stress variations within plant canopies.
• Variations in chlorophyll content and stress levels across leaves limit the accuracy of traditional phenotyping.

Purpose of the Study:

• To develop a novel leaf image processing algorithm for analyzing nutrient and stress distribution patterns.
• To improve the quality and accuracy of plant phenotyping measurements.
• To demonstrate the algorithm's effectiveness in differentiating nitrogen stress levels and genotypes in corn.

Main Methods:

• Developed a new leaf image processing algorithm integrating Random Forest and leaf region rescaling.
• Utilized the normalized difference vegetation index (NDVI) to assess algorithm performance.
• Modeled distribution patterns along the corn leaf's mid-rib direction.

Main Results:

• The new algorithm successfully modeled and utilized distribution patterns for enhanced phenotyping.
• It clearly differentiated leaves based on nitrogen treatments and genotypes.
• The algorithm showed improved accuracy compared to traditional methods that average NDVI across the whole leaf.

Conclusions:

• The developed algorithm offers a finer level of image analysis for plant phenotyping.
• Analyzing nutrient and stress distribution patterns provides valuable signals for improved crop assessment.
• This approach holds potential for enhancing various plant feature measurements beyond NDVI.