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.

Feature Wavelengths Selection Using Successive Projections Algorithm for Prediction of Apple Firmness and Soluble Solids Content

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

Citation:  2010 Pittsburgh, Pennsylvania, June 20 - June 23, 2010  1008913.(doi:10.13031/2013.29755)
Authors:   Min Huang, Renfu Lu
Keywords:   Hyperspectral scattering, Hybrid successive projections algorithm, Kennard-Stone algorithm, Multiple linear regression, Fruit, Apple, Firmness, Soluble solids content.

Hyperspectral scattering is a promising technique for noninvasive measurement of quality attributes of apple fruit. A hybrid successive projections algorithm (HSPA) coupled with multiple linear regression (MLR) was proposed to extract the feature wavelengths from the hyperspectral scattering profiles of Golden Delicious apples for predicting fruit firmness and soluble solids content (SSC). Six hundred samples were tested in the experiment, 400 of which were used to develop calibration models with feature wavelengths selected by the Kennard-Stone algorithm and the remaining 200 fruits were used for validation. Eleven feature wavelengths were selected for firmness, which nearly spanned the entire spectral range of 500 - 1,000 nm, and 18 feature wavelengths, including one below 600 nm, were selected in the SSC prediction model. The model using feature wavelengths for predicting firmness yielded better results (root mean squared error of prediction or RMSEP = 6.1 N) than the MLR models using wavelengths selected by forward selection (FS) and successive projections algorithm (SPA). For predicting SSC, the result for the model using 18 feature wavelengths selected by the HSPA method was mixed compared with the FS-MLR and SPA-MLR models. Upon further test and validation, the proposed HSPA approach could be suitable for feature wavelengths selection for prediction of firmness and SSC in apples.

(Download PDF)    (Export to EndNotes)