[Study on method of maize hybrid purity identification based on hyperspectral image technology]

Area of Science:

• Agricultural Science
• Spectroscopy
• Image Analysis

Background:

• Maize hybrid purity is crucial for seed production and crop performance.
• Accurate identification methods are needed to ensure seed quality and prevent misrepresentation.

Purpose of the Study:

• To evaluate the feasibility of hyperspectral imaging for identifying maize hybrid purity.
• To develop and validate models for distinguishing maize hybrids from their female parents using spectral data.

Main Methods:

• Collected hyperspectral images (871-1699 nm, 308 wavelengths) of maize variety NH101 seeds.
• Extracted spectral information and built identification models.
• Investigated the impact of sample orientation and environmental factors on model performance.
• Identified key spectral bands (1195-1246 nm) for differentiation using the Qs method.

Main Results:

• Identification models achieved average correct acceptance and rejection rates exceeding 90%.
• Model performance was robust across different sample laying modes and experimental environments.
• Models built using selected feature bands performed comparably to those using the full spectral range.

Conclusions:

• Hyperspectral imaging technology is a feasible, objective, and rapid method for identifying maize hybrid purity.
• The identified spectral bands offer a targeted approach for efficient maize hybrid identification.