<![CDATA[Cavendish banana (Musa acuminata Cavendish) is one of the horticultural crops whose ripening is often accelerated using calcium carbide (CaC₂), which is harmful to health, thus requiring a scientific method to distinguish natural and artificial ripening. This study aimed to examine the potential of Visible–Near Infrared (Vis-NIR) spectroscopy using the AS7265X multispectral sensor, operating at 410–940 nm, which is portable and more affordable than a laboratory spectrophotometer. A total of 120 samples were used, consisting of 90 unripe (stage 2) and 30 naturally ripened (stage 6) bananas. Linear Discriminant Analysis (LDA) was employed to classify the spectral data, achieving a classification accuracy of 100%. The Vis-NIR spectral patterns showed apparent differences among treatments. Unripe bananas had high reflectance in the blue–green region, while tree-ripened bananas showed increased reflectance in the red and NIR regions. The 64 g/kg carbide treatment yielded a spectral pattern resembling natural ripening, whereas the single lump carbide treatment showed lower reflectance values across most wavelengths. These findings confirmed the potential of the AS7265X sensor to efficiently and non-destructively distinguish between natural and artificial ripening. Practically, this suggests that low-cost, portable sensors can be effectively deployed for real-time field inspection and quality control within the fruit supply chain. Future study need to validate the method using larger and independent datasets.]]>
Copyrights © 2026
