Journal of Food Processing & Technology

Determination of weight and color in Agraz during the storage period with near infrared spectroscopy

19^th International Conference on Food Processing & Technology

October 23-25, 2017 | Paris, France

Andrea Katherin Carranza Diaz, Nathalia Forero-Cabrera, Yelena Gomez-Lara, Laura Mila-Saavedra, Andres Silva-Dimate and Carolina Sanchez-Saenz

National University of Colombia, Colombia

Scientific Tracks Abstracts: J Food Process Technol

Abstract:

64% of fruit and vegetable losses occur in the post-harvest and industrial processing stages. To determine the quality of the product, destructive methods are frequently used. Near-infrared spectroscopy (NIR) is a technology considered as a nondestructive alternative for the determination of quality and maturation indices. Agraz (Vaccinium meridionale Swartz) known as Colombian wild blueberry, is a plant belonging to the family Ericaceae, considered an essential fruit for its antioxidant values, however, no spectroscopic studies are reported for determination of quality parameters. The aim of this work was to calibrate and validate models for prediction of weight and color in agraz during its storage period with NIR spectroscopy. Vegetative material came from the municipality of San Miguel de Sema, Boyacá, Colombia; agraz was harvested at maturity stage 3 and selected according to its phytosanitary state. It was stored for 21 days at room temperature (18°C±0.5 and 60% RH±0.5) and weight changes and color (CIELAB L* a* b*) of 100 samples of 15 g±0.2 g of product were monitored. 300 samples for calibration and 100 for external validation were used. For the elaboration of the models, PLSR was used and cross-validation was performed and for their evaluation R2, RMSE, RPD were considered as criteria. Good representation models were obtained for weight an R2 of 0.76 and 0.072 RPD. For color, acceptable models, a* coordinate with RPD 1.53 for b* coordinate models rated average prediction (R2 0.41) and in the case of the coordinate L* is obtained a model with poor predictive capacity. The results indicate that the spectral models can be used as tools of classification and measurement of weight and color with some precision in the coordinates a* and b*, however, it is necessary to improve the models.

Biography :

Andrea Katherin Carranza Diaz is a Bachelor student from National University of Colombia, Colombia. Her research interests include Postharvest of agricultural products and soil sciences.