Hyperspectral imaging and predictive microbiology for the non-invasive evaluation of the growth probability of Staphylococcus aureus in sliced dry-cured ham.

This work aimed to explore the potential of a non-invasive approach, consisting of the use of hyperspectral imaging (HSI) together with predictive microbiology, for dry-cured ham safety evaluation. The growth probability of Staphylococcus aureus in salt-reduced or not sliced dry-cured ham stored at room temperature and the effect of image preprocessing on this evaluation were analysed. To this end, predictive models for a were developed (RMSEP = 0.013) and later used to calculate a chemical images, a frequency polygons and predicted a at different percentiles, which were later used as predictive microbiology model inputs. This innovative approach showed no differences in product a between salting groups and in the associated growth probability of S. aureus. Due to the high variability of a between samples inside the same salting batch, the growth probability of S. aureus ranged from 19 to 46 % when using pH and a values at percentile 75th. Image preprocessing was able to remove image artifacts (specular highlights and fat streaks), and the different image preprocessing thresholds used did not influence the predicted a values at different percentiles.