Identification of Haploid Maize Kernel Using NIR Spectroscopy in Reflectance and Transmittance Modes: A Comparative...

Area of Science:

• Agricultural Science
• Spectroscopy
• Biotechnology

Background:

• Haploid maize production is crucial for breeding, but efficient identification of haploid kernels is challenging.
• Traditional methods for haploid identification are often labor-intensive and time-consuming.

Purpose of the Study:

• To explore suitable spectra measurement modes for accurate haploid maize kernel identification.
• To compare the effectiveness of reflectance and transmittance Near Infrared Spectroscopy (NIRS) for this purpose.

Main Methods:

• Utilized a miniature near-infrared spectrometer (MicroNIR-1700) for spectral measurements.
• Applied Near Infrared Spectroscopy (NIRS) qualitative analysis techniques.
• Employed Partial Least Squares-Discriminant Analysis (PLS-LDA) for data compression and Support Vector Machine (SVM) for model building.
• Compared reflectance and transmittance spectral data for haploid kernel identification.

Main Results:

• Reflectance mode yielded an average recognition rate below 60%, highly sensitive to embryo side position.
• Transmittance mode achieved a significantly higher average recognition rate of 93.2%.
• Transmittance mode proved effective and less sensitive to kernel orientation, capturing interior information.

Conclusions:

• Near-infrared diffuse transmittance spectroscopy is a highly effective, rapid, and nondestructive method for identifying haploid maize kernels.
• This technique offers advantages over reflectance mode by providing comprehensive internal sample information.
• The Micro-NIR spectrometer's speed and cost-effectiveness support automated haploid maize selection.