TY - JOUR

T1 - Transferring results from NIR-hyperspectral to NIR-multispectral imaging systems

T2 - A filter-based simulation applied to the classification of Arabica and Robusta green coffee

AU - Calvini, Rosalba

AU - Amigo Rubio, Jose Manuel

AU - Ulrici, Alessandro

PY - 2017

Y1 - 2017

N2 - Due to the differences in terms of both price and quality, the availability of effective instrumentation to discriminate between Arabica and Robusta coffee is extremely important. To this aim, the use of multispectral imaging systems could provide reliable and accurate real-time monitoring at relatively low costs. However, in practice the implementation of multispectral imaging systems is not straightforward: the present work investigates this issue, starting from the outcome of variable selection performed using a hyperspectral system. Multispectral data were simulated considering four commercially available filters matching the selected spectral regions, and used to calculate multivariate classification models with Partial Least Squares-Discriminant Analysis (PLS-DA) and sparse PLS-DA. Proper strategies for the definition of the training set and the selection of the most effective combinations of spectral channels led to satisfactory classification performances (100% classification efficiency in prediction of the test set).

AB - Due to the differences in terms of both price and quality, the availability of effective instrumentation to discriminate between Arabica and Robusta coffee is extremely important. To this aim, the use of multispectral imaging systems could provide reliable and accurate real-time monitoring at relatively low costs. However, in practice the implementation of multispectral imaging systems is not straightforward: the present work investigates this issue, starting from the outcome of variable selection performed using a hyperspectral system. Multispectral data were simulated considering four commercially available filters matching the selected spectral regions, and used to calculate multivariate classification models with Partial Least Squares-Discriminant Analysis (PLS-DA) and sparse PLS-DA. Proper strategies for the definition of the training set and the selection of the most effective combinations of spectral channels led to satisfactory classification performances (100% classification efficiency in prediction of the test set).

KW - Green coffee

KW - Hyperspectral imaging

KW - Multispectral imaging

KW - Multivariate classification

KW - Sparse methods

U2 - 10.1016/j.aca.2017.03.011

DO - 10.1016/j.aca.2017.03.011

M3 - Journal article

C2 - 28390483

AN - SCOPUS:85016080005

VL - 967

SP - 33

EP - 41

JO - Analytica Chimica Acta

JF - Analytica Chimica Acta

SN - 0003-2670

ER -