Articles
Development of predictive models for fruit quality with portable NIR spectroscopy evaluation in pitaya produced in Mediterranean greenhouses
Article number
1452_22
Pages
171 – 178
Language
English
Abstract
Protected cultivation of pitaya (Selenicereus undatus) represents a promising opportunity for growers in the Mediterranean region.
To ensure its success and profitability, it is essential to ascertain specific fruit quality parameters at harvest, which facilitates its adequate subsequent handling for commercialization and optimum quality for consumers.
In this context, most of the physico-chemical parameters that determine fruit quality are traditionally evaluated by destructive analyses, which implies significant time and high economic impacts.
To offer an alternative that does not result in a loss of yields and saves time and costs, this study aimed to develop and validate predictive models using portable near-infrared spectroscopy (NIR) to determine skin color, total soluble solids content (TSS) and acidity in pitaya fruits, as a non-destructive way, and to establish their relationship with its maturity stage at harvest.
For this, a total of 164 fruits at three maturity levels (early, medium and late) were harvested in a southeastern Spain greenhouse and evaluated at the laboratory, using a portable VIS-NIR spectrometer (F-750, Felix Instrument, Camas, USA). The data set was divided into two subsets: one for calibration and the other for model validation.
Based on the data, some spectrum pre-processing techniques, wavelength selection, and discriminant methods, were utilized in model generation.
Subsequently, nonlinear modeling techniques were analyzed that afforded several divergent models.
In the best ones, a coefficient of determination (R2) near 0.9 and a root-mean-square error (RMSE) of less than 1 for all parameters under study were reached.
Therefore, the application of portable NIR in combination with chemometric techniques can be employed in the prediction of skin color, TSS and acidity in pitaya, as well as the determination of the maturity stage of fruits.
To ensure its success and profitability, it is essential to ascertain specific fruit quality parameters at harvest, which facilitates its adequate subsequent handling for commercialization and optimum quality for consumers.
In this context, most of the physico-chemical parameters that determine fruit quality are traditionally evaluated by destructive analyses, which implies significant time and high economic impacts.
To offer an alternative that does not result in a loss of yields and saves time and costs, this study aimed to develop and validate predictive models using portable near-infrared spectroscopy (NIR) to determine skin color, total soluble solids content (TSS) and acidity in pitaya fruits, as a non-destructive way, and to establish their relationship with its maturity stage at harvest.
For this, a total of 164 fruits at three maturity levels (early, medium and late) were harvested in a southeastern Spain greenhouse and evaluated at the laboratory, using a portable VIS-NIR spectrometer (F-750, Felix Instrument, Camas, USA). The data set was divided into two subsets: one for calibration and the other for model validation.
Based on the data, some spectrum pre-processing techniques, wavelength selection, and discriminant methods, were utilized in model generation.
Subsequently, nonlinear modeling techniques were analyzed that afforded several divergent models.
In the best ones, a coefficient of determination (R2) near 0.9 and a root-mean-square error (RMSE) of less than 1 for all parameters under study were reached.
Therefore, the application of portable NIR in combination with chemometric techniques can be employed in the prediction of skin color, TSS and acidity in pitaya, as well as the determination of the maturity stage of fruits.
Authors
I. Portillo, I. Salinas, V. Pinillos
Keywords
Selenicereus undatus, protected cultivation, near-infrared spectroscopy, chemometric techniques, mathematical modeling, fruit attributes, fruit maturity
Groups involved
Online Articles (65)