Garlic bulb classification by combining Raman spectroscopy and machine learning

The purpose of this study was to demonstrate the utility of combining Raman spectroscopy with machine learning techniques for achieving origin traceability of five garlic bulb species. We collected Raman spectra of garlic bulbs and Raman bands are assigned. After pre-processing, the wavenumbers and intensities of distinct Raman peaks are extracted as the input data for developing the classification model. Our trained model presents an accuracy of 98.97%, a precision of 98.92% and a sensitivity of 98.86%. The results indicate that the artificial prior feature extraction strategy prevents over-fitting due to external variables and improves greatly model accuracy. This study offers a novel classification and origin identification scheme for plant bulbs.

Automated summary

The study demonstrates the combination of Raman spectroscopy and machine learning for traceability of garlic bulb species. Raman spectra of garlic bulbs were collected and analyzed for developing a classification model. The trained model achieved high accuracy, precision, and sensitivity. The study offers a novel approach for classification and origin identification of plant bulbs.