Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Near-infrared Spectroscopy to Predict Soluble Solids and Firmness in Apples

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  Paper number  023066,  2002 ASAE Annual Meeting . (doi: 10.13031/2013.9628) @2002
Authors:   Bosoon Park, Judith A. Abbott, Kang-Jin Lee, Chang-Hyun Choi, Kyu-Hong Choi
Keywords:   Near-infrared, Principal Component Regression (PCR), Nondestructive, Fruit, Sugar Contents, Mahalanobis Distance (MD), Soluble Solids, Firmness, Spectroscopy

Development of nondestructive measurements of soluble solids and firmness of apples benefits the producers, processors, and packers. The soluble solids and Magness-Taylor firmness of apples were predicted by spectroscopic diffuse reflectance measurement at the visible/near-infrared (400-2500 nm) wavelengths. Two cultivars, Pennsylvania Gala (n=800) and Washington Delicious (n=960) apples, were measured. The optical measurement of soluble solids and firmness for each apple were validated by a refractometer and a universal testing instrument (Instron), respectively. Based on the optical properties of apples, the soluble solids could be predicted by NIR spectroscopic techniques with principal component regression (PCR) and Mahalanobis distance (MD) analysis. The coefficients of determination (R2) for prediction of soluble solids were 0.934 (SEP=0.279; 10 factors) for Gala apples; and 0.966 (SEP=0.341; 10 factors) for Delicious apples with a NIR region of spectrum (800-1100 nm). For classifying Gala apples into three classes based on the soluble solids, MD classifiers had classification accuracies of 93.5% for the full spectrum range (400-1800 nm) and 95.5% for the partial NIR spectrum range (800-1100 nm). Similarly, for the classification of Delicious apples, the classification accuracies were 92.1% with the full spectrum (400-1800 nm) and 93.6% with the partial NIR spectrum (800-1100 nm). A spectroscopic technique for apple firmness measurements was feasible for Delicious apples. Using PCR models, the coefficients of determination (R2) for predicting firmness (Magness-Taylor maximum force) were 0.218 (SEP=4.91; 16 factors) for Gala apples and 0.786 (SEP=7.02; 24 factors) for Delicious apples with full spectrum (400-1,800 nm); while, the R2s were only 0.291 (SEP=4.92; 12 factors) for Gala and 0.657 for Delicious (SEP=7.33; 24 factors) with the partial NIR spectrum (800-1100 nm). The MD analyses were also conducted to classify apples firmness. Based on three classes, the classification accuracies were 82.5% with the full range of spectrum (400-1800 nm); and 80% with partial NIR spectrum (800-1100 nm) for Gala apples. Similarly, the classification accuracies were 83.8% with the full spectrum (400-1800 nm); and 75.3% with the partial NIR spectrum (800-1100 nm) for Delicious apples. This shows that spectroscopic techniques were feasible to classify apple firmness with over 82% accuracy.

(Download PDF)    (Export to EndNotes)