This paper (open access) proposes an approach to deal with surface scattering in the Near-Infrared (NIR) analysis of particulates. The authors use coffee cultivarl testing as an example.
Surface scattering is a major confound in near-infrared (NIR) analysis of particulate foods. In roasted coffee powders, inhomogeneous scattering can obscure cultivar differences. Standard practice eliminates scattering (extended/multiplicative scatter correction) via reference-anchored polynomial projection. The problem is that in heterogeneous matrices this is order-sensitive and can remove analyte-relevant variance.
The approach proposed in this paper is to encode scattering explicitly. Per-spectrum polynomial slope, curvature and cubic baseline coefficients are determined and appended as descriptors and models are trained on the augmented matrix.
The authors reported that, using 300 diffuse-reflectance FT-NIR spectra (10 000–4000 cm−1) from 25 lots covering 7 Arabica cultivars, this strategy improved test-set authentication. Linear Discriminant Analysis (LDA) increased from 70.8% to 83.3%; Support Vector Machine (SVM)-linear from 63.9% to 84.7%, while Quadratic Discriminant Analysis (QDA) remained high (88.9%).
They conclude that the approach provides a quantitative scattering-aware method, treating it as information rather than aiming for its blind elimination. This method is an interpretable, easily implemented alternative and is applicable to other particulate and microstructured foods (e.g., cocoa, tea, spices)
Comments