Feasibility of Vis/NIR spectroscopy and image analysis as basis of the development of smart-drying technologies

Abstract

Drying is a complex, dynamic, unsteady and nonlinear process that, when not optimized on a system level, may be responsible for (1) significant quality degradation and (2) energy wastage. Consequently, new drying technologies must be designed combining non-invasive at-/on-/in-line advanced measurement and control systems with models cross-linking all relevant aspects of product quality changes and heat and mass transfer phenomena. This paper presents preliminary results on the use of RGB imaging, NIR spectroscopy and Vis-NIR hyperspectral imaging for real-time monitoring of physicochemical changes of apples and carrots during drying. Keywords: chemometrics, artificial intelligence, deep learning