Non-destructive detection of specific spoilage bacteria growth and sensory change on chilled pork surface by Vis/NIR spectroscopy

A rapid technique was applied to assess chilled pork spoilage employing a non-destructive visible and near-infrared (Vis/NIR) spectroscopy method. As specific spoilage organism (SSO), Pseudomonas spp. played a dominant role in causing chilled meat spoilage. Microbial growth and sensory change based on two indexes (Pseudomonas spp. enumeration and color) were determined to appraise pork spoilage. A total of 45 samples were packed in sterilized plastic pallet and stored at 4°C for 1 to 15 days. Hyperspectral images in the range of 380 to 1100 nm were extracted from meat surface. Standard normal variables (SNV) and mean filtering (MF) pre-treatment methods were employed to eliminate the spectra noise. Then principal component analysis (PCA) was used to reduce the date dimensionality. The support vector machines (SVM) was applied to establish prediction models. The models yielded regression coefficient of prediction set (Rp) were 0.95 and 0.94 for Pseudomonas spp. enumeration and color respectively. Classification models using Bayesian discriminant analysis were developed to classify the pork quality into three categories of grades. All samples then were classified in fresh grade with 100% accuracy, fresh to spoilage grade with 80% and spoilage with 100% accuracy. The study demonstrated that Vis/NIR spectroscopy technique combined with Bayesian discriminant analysis was a precise and potential tool for assessing microbiological spoilage of chilled pork.