Comparison of different illumination systems for moisture prediction in cereal bars using hyperspectral imaging technology

Moisture content and its distribution is a critical parameter in the production of cereal bars. Inappropriate control this quality can lead to non-conforming products excess waste on lines. In field hyperspectral imaging, search for alternative light sources stabilised-halogen (cheaper emitting less heat) growing need application technology industry. This study compares three different illumination systems moisture prediction visible-near infrared (vis-NIR) range (from 400 nm 1000 nm). The images were acquired using including two halogen-based (stabilised-halogen conventional-halogen) an LED-based system. results showed that combined with partial least squares model better predicted Lower accuracies obtained when experiment was performed system, which double error systems. It concluded consequence information lost bands appearing above 850 may be revealing about bars since second overtone water O–H found at 970 nm. demonstrate conventional vis-NIR are promising method