Prediction of Starch, Soluble Sugars and Amino Acids in Potatoes (Solanum tuberosum L.) Using Hyperspectral Imaging, Dielectric and LF-NMR Methodologies
- 458 Downloads
Handling and processing of potatoes is performed in increasingly large and more automated facilities, and the industry calls for more automated machinery for quality assessment and sorting by concentration of starch, soluble sugars, protein, amino acids etc. of the potato tubers. The present study was designed to evaluate five different scanning methods for their potential use in potato assessment and sorting. Two methods were based on hyperspectral imaging, two were based on dielectric/bio-impedance and one was based on low-field nuclear magnetic resonance. A set of 60 potatoes of 10 different cultivars were simultaneously sampled for analyses of content and scanned by the five different scanning methods. The resulting multivariate dataset was used to estimate the prediction ability of the individual scanning methods on starch-related parameters, selected simple sugars, selected amino acids, conductivity of pressed cell sap and cell sizes. Results showed that most types of spectral analyses had relatively high potential for predicting the starch-related parameters and medium potential for predicting the concentration of the reducing sugars fructose and glucose. Most methods showed medium potential for prediction of several amino acids, including asparagine, which showed particularly promising predictions in the hyperspectral analyses of intact potatoes. The presented screening study enabled us to perform robust choices for the further development and optimization of the methods and instruments for industrial implementation.
KeywordsAmino acids Dielectric function Hyperspectral imaging LF-NMR Starch Sugars
The present study was funded by the Innovationsfonden, Denmark [129-2013-5], Newtec Engineering and Aarhus University. The authors wish to extend their gratitude to Bjarne Thiesgaard, AKS and Ruth Madsen, Danespo A/S, for supplying the potato samples and to laboratory technicians Jens Madsen, Nina Eggers, Annette Brandsholm, Karin Henriksen and Elmedina Dervisevic for their efforts in the project.
- Breiman L, Friedman J, Stone CJ, Olshen RA (1984) Classification and regression trees. The Wadsworth and Brooks-Cole statistics-probability series. Taylor & Francis, New YorkGoogle Scholar
- Dacal-nieto A, Formella A, Carrion P, Vazquez-Fernandez E, Fernandez-Delgado M (2011) Non-destructive detection of hollow heart in potatoes using hyperspectral imaging. Comp Sci Theory Methods 6855:180–187Google Scholar
- Gamer M, Lemon J, Fellows I (2012) The irr package: various coefficients of interrater reliability and agreement. R package version 0.84Google Scholar
- Kuhn M, Wing J, Weston S et al (2015) The caret package: classification and regression training. R package version 6.0-47. http://CRAN.R-project.org/package=caret
- Mevik B-H, Wehrens R, Liland KH (2013) The pls package: principal component and partial least squares regression in R. PLS Partial Least Squares Princ. Compon. regression. R Package. version 2.4-3. http://CRAN.R-project.org/package=pls
- Mizukami Y, Yamada K, Sawai Y, Yamaguchi Y (2007) Measurements of fresh tea leaf growth using electrical impedance spectroscopy. Agric Journal, Medwell 2:134–139. doi:aj.2007.134.139Google Scholar
- R Development Core Team (2015). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org
- Rady A, Guyer D, Lu R (2015) Evaluation of sugar content of potatoes using hyperspectral imaging. Food Bioprocess Technol 995–1010. doi: 10.1007/s11947-014-1461-0
- Stone M (1974) Cross-validatory choice and assessment of statistical predictions. J R Stat Soc 36:111–147Google Scholar