Use of expert knowledge for improved quality monitoring of food colloids with Vis/NIR spectroscopic techniques.
Quality monitoring of food colloids during production and processing is important to satisfy the increasing quality demands of the retail and consumers. Moreover, online measurement techniques could support early detection of altering product quality and reduce food losses. Optical spectroscopic techniques measure the interaction of light with the product. As the measured signal(s) contain both chemical and physical product information, while the techniques are generally non-destructive and do not require extensive sample preparation, they are very promising for quality monitoring in the agrofood sector. The measured optical signals are, however, the result of the complex interactions between light absorption and scattering, respectively caused by the chemical constituents and the local nonuniformities (physical properties) in the measured food product. To extract unambiguous information on the sample composition and physical microstructure (particle size distribution and quantity, …) from these spectroscopic measurements, the measurements have to be combined with a model that mathematically describes light propagation. In this study, we aim to increase the robustness and efficiency of this light propagation model through incorporating expert knowledge on the relation between scattering and the physical microstructure typical for food colloids.