Raman and FT-IR Spectroscopy Coupled with Machine Learning for the Discrimination of Different Vegetable Crop Seed...

Area of Science:

• Spectroscopy
• Chemometrics
• Bioinformatics

Background:

• Accurate seed variety identification is crucial for crop management and genetic resource preservation.
• Traditional methods can be time-consuming and labor-intensive.

Purpose of the Study:

• To evaluate Raman and FT-IR spectroscopy for discriminating between seed varieties of paprika, tomato, and lettuce.
• To assess the efficacy of various chemometric models and machine learning algorithms for spectral data analysis.

Main Methods:

• Seeds were analyzed using Raman and FT-IR spectroscopy.
• Spectral data underwent pre-processing including smoothing, baseline correction, and normalization.
• Classification was performed using Principal Component Analysis (PCA), Support Vector Machines (SVM), Partial Least Square Discriminant Analysis (PLS-DA), and PCA-Quadratic Discriminant Analysis (PCA-QDA).

Main Results:

• Support Vector Machines (SVM) achieved high classification accuracies: 100% for lettuce, 99.37% for paprika, and 92.71% for tomato using Raman spectroscopy.
• FT-IR spectroscopy with SVM yielded accuracies of 99.37% for lettuce, 92.50% for paprika, and 97.50% for tomato.
• Merging Raman and FT-IR spectra improved classification accuracy, reaching 100% for lettuce and tomato, and 95% for paprika.

Conclusions:

• Raman and FT-IR spectroscopy, coupled with machine learning, provide a rapid and effective method for seed variety discrimination.
• This technique holds significant potential for the evaluation and management of genetic resources in seed banks.