In this study (open access) the authors made a reference dataset of comminuted meat mixtures by dicing and mixing 140 commercially-purchased steaks of beef, duck and chicken. They built a classification model to discriminate between the three species in the mixtures.
They used a hand-held Hyperspectral Imaging (HSI) (with a Raspberry Pi controller, which has real-time image acquisition and processing covering a spectral range from 400 nm to 800 nm) to develop a discrimination model for chicken/duck adulteration in diced beef. The portable push broom HSI was designed with the spectral resolution of 5 nm and spatial resolution of 0.1 mm. To improve generalization, a model transfer method was also developed to achieve model sharing across instruments
The authors report that their model transfer method can effectively correct the spectral differences due to instrument variation and improve the robustness of the model. The support vector machine (SVM) classifier combined with spectral space transformation (SST) achieved a best accuracy of 94.91%. Additionally, a visualization map was proposed to provide the distribution of meat adulteration.
They conclude that the portable HSI enables on-site analysis, making it an invaluable tool for various industries, including food safety and quality control.
Comments