An efficient method for the rapid detection of industrial paraffin contamination levels in rice based on hyperspectral imaging

Abstract

The addition of industrial paraffin to rice is an important food safety issue, which brings potential risks to human health. Therefore, an efficient and rapid method for detecting industrial paraffin in rice is an urgent requirement. In this study, rice samples from six regions in China were tested by hyperspectral imaging to detect industrial paraffin contamination levels (IPCLs). The local model predicted the IPCL using a local rice training dataset, and obtained excellent prediction results; however, the cross-transitivity analysis of these models showed that they can only be applied to local regions’ datasets. Further, the transfer of local models can help PLSR learn relevant features by transferring samples from the target dataset to be tested to the source dataset, and only required 40% of the off-site samples to supplement the source dataset to achieve an effective assessment of IPCL. Furthermore, the multi-source model is suitable for predicting the IPCL of rice in different regions simultaneously. Finally, the competitive adaptive reweighted sampling (CARS) method is used to optimize the three aforementioned models. The optimized average correlation coefficients R are 0.9631, 0.8805, and 0.8542, respectively, and the optimized average RMSE are 0.067%, 0.121%, and 0.132%, respectively.