Towards the development of a sterile model cheese for assessing the potential of hyperspectral imaging as a non-destructive fungal detection method

Abstract

Food can be spoiled by microbial contaminants rendering it unsafe for human consumption, and requires regular testing to ensure its safety. Current techniques involve the destruction of food matrices for microbial assessment. In contrast, emerging technologies such as hyperspectral imaging (HSI) can be integrated into the manufacturing chain as on-line or in-line systems for frequent microbial assessment. This study aims to develop a model cheese system with a high surface area to optimise and deploy HSI for microbial detection. Compatibility and stability studies were conducted on the model cheese system by measuring physiochemical parameters, such as pH moisture content, water activity, total protein content, and sugars. Microbial testing was also carried out to quantify the natural or background microflora of the model cheese after rennet and Lactobacillus spp. starter cultures were added. Penicillium chrysogenum was isolated from commercial cheese samples using 18S rRNA gene amplification and sequencing. Finally, the model cheese was challenge tested with P. chrysogenum isolated from commercial samples and imaged using a hyperspectral camera. As a proof of concept, an untrained principal component analysis model was developed to identify model cheese inoculated with P. chrysogenum at mycelial growth or after conidia formation, illustrating a versatile model cheese system that can be deployed for non-destructive assessment of fungal contaminants with HSI.