Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
6665579 | Journal of Food Engineering | 2015 | 6 Pages |
Abstract
Before pineapples are canned, the ones with high nitrate level must be sorted out first because nitrate causes black stains on the surface of the can; therefore, a nondestructive technique for sorting out pineapples is clearly needed. The use of visible and near infrared (Vis-NIR) spectroscopy for such purpose was investigated in this study. A batch of 75 pineapple fruits that would have been delivered to a canning factory was tested. Spectra were acquired using a spectrophotometer in interactance mode with wavelengths in the region of 400-2500Â nm. Twelve scans of different parts of each pineapple were made. The actual amount of nitrate in the pineapple flesh was determined by HPLC. Original spectra and pretreated spectra were both used to construct calibration models with partial least squares regression (PLSR). The best model was obtained from an average spectrum pretreated with first derivative treatment at the wavelength range of 600-1200Â nm. Predictions based on this model matched closely with the actual nitrate contents, with a high correlation coefficient (R) of 0.95 and a low root mean square error of prediction (RMSEP) of 1.77Â ppm. These results demonstrate that Vis-NIR spectroscopy can be used for rough screening of intact pineapple.
Keywords
Related Topics
Physical Sciences and Engineering
Chemical Engineering
Chemical Engineering (General)
Authors
Sasathorn Srivichien, Anupun Terdwongworakul, Sontisuk Teerachaichayut,