Publication | Closed Access
NON-DESTRUCTIVE MEASUREMENT OF ACIDITY, SOLUBLE SOLIDS, AND FIRMNESS OF JONAGOLD APPLES USING NIR-SPECTROSCOPY
273
Citations
0
References
1998
Year
EngineeringFood AnalysisMechanical EngineeringAgricultural EconomicsRipeningCrop QualityFood ChemistrySoluble SolidsOptical PropertiesCalibrationAnalytical ChemistryBiostatisticsFruit SciencePrincipal Component AnalysisPost-harvest PhysiologyBiophysicsHorticultural ScienceHealth SciencesBifurcated Optical FibreJonagold ApplesNear-infrared SpectroscopyFood QualitySpectroscopy
The study evaluates visible and near‑infrared spectroscopy (380–1650 nm) as a non‑destructive method to assess quality traits of Jonagold apples. Reflectance spectra collected with a bifurcated optical fibre were linked to pH, soluble solids, stiffness, and texture, and multivariate calibration (PCA, PCR, PLS) with preprocessing was used to build predictive models. The models achieved good to excellent accuracy, with the pH model yielding SEP 0.068 (r 0.93), stiffness SEP 2.49 (r 0.90), soluble solids SEP 0.61 (r 0.82), and elastic modulus SEP 0.26 (r 0.75), indicating a strong physico‑chemical basis for pH but limited performance for elasticity.
In this study the potential of visible and near infrared spectroscopy (380-1650 nm) as a non-destructivemeasurement technique for measuring quality characteristics of Jonagold apples is evaluated. A relation has beenestablished between the reflectance spectra recorded with a bifurcated optical fibre and fruit parameters such as the pH,the soluble solids content, the stiffness factor and other texture parameters such as the elastic modulus of the flesh. Thedata were analyzed with multivariate calibration techniques such as principal component analysis (PCA), principalcomponent regression (PCR) and partial least squares analysis (PLS). The impact of the data preprocessing on theprediction performance has also been investigated. The prediction models indicate a good to excellent predictionperformance. The best pH model has a standard error of prediction (SEP) of 0.068 and a correlation between predictedand measured values (r) of 0.93. The proposed model for the stiffness factor has a SEP of 2.49 and an r-value of 0.90. Forthe soluble solids content model a SEP of 0.61 and an r-value of 0.82 is obtained. It has been shown that the pH modelhas a physico-chemical background. The prediction performance of the model for the elastic modulus, however, was notyet satisfactory (SEP = 0.26, r = 0.75).