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.

Determination of soluble solids content in apple using hyperspectral imaging and variable selection algorithms

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

Citation:  Paper number  131620975,  2013 Kansas City, Missouri, July 21 - July 24, 2013. (doi: @2013
Authors:   Wenqian Huang, Liping Chen, Jiangbo Li, Zhiming Guo
Keywords:   Hyperspectral imaing apple soluble solids content ant colony optimization successive projections algorithm multivariate calibration analysis.

Abstract. Today, determination of soluble solids content in apple based on hyperspectral imaging is slow for the long acquisition time of hyperspectral images. Therefore it is important to select effective wavelengths from hyperspectral data to decrease the acquisition time. In this study, ant colony optimization (ACO) and successive projections algorithm (SPA) were used for extracting effective wavelengths from 769 wavelengths of 400~1 000 nm hyperspectral reflectance images of ‘Fuji’ apples. A total of 160 apple samples were prepared for the calibration (n=120) and prediction (n=40) sets. Different preprocessing methods were compared and the autoscaling was determined as the best one. Based on the 51 and 39 effective wavelengths selected by ACO and SPA respectively, different models were built and compared for predicting soluble solids content (SSC) in apple using partial least squares (PLS) and multiple linear regression (MLR). Within all the models, the SPA-MLR achieved the best results, where , and were 0.9, 0.3 and 4. respectively. However, , and of the ACO-MLR model were 0.8046, 0.6921 and 2.2622 respectively, which was not a good results compared with these of the PLS model based on 769 wavelengths. Results showed that SPA could be used for selecting the effective wavelengths from hyperspectral data. The SPA-MLR is an optimal modeling method for prediction of SSC in apple and has a great potential for on-line detection of SSC in apple.

(Download PDF)    (Export to EndNotes)