Spectral imaging for contamination detection in food

Spectral imaging is a technique with a big potential for surface chemistry mapping of heterogeneous samples. It works by making a spectrum in every pixel of an image, and this spectrum may under the right circumstances be transformed into abundance maps for chemical components. One important application of the technique is finding anomalies I supposedly homogeneous matter or homogeneous mixtures. This application occurs frequently in the food industry when different types of contamination are to be detected. Contaminants could be e.g. foreign matter, process-induced toxins, and microbiological spoilage. Many of these contaminants may be detected in the wavelength range visible to normal silicium-based camera sensors i.e. 350-1050 nm with proper care during sample preparation, sample presentation, image acquisition and analysis. This presentation will give an introduction to the techniques behind the VideometerLab instrument, that implements the thoughts above, and show examples including fusarium detection in barley, measuring microbial meat spoilage, and making humidity maps. It will also illustrate methodology for spectral image analysis.