Abstract: Terahertz time-domain spectroscopy is a kind of far-infrared spectroscopy technology, and its spectrum reflects the internal properties of substances with rich physical and chemical information, so the use of terahertz waves can be used to qualitatively identify food additives containing nitrogen elements. Analytic hierarchy process (AHP) was originally used to solve evaluation-type problems, and this paper introduces it into the field of terahertz spectral qualitative analysis, proposes a terahertz time-domain spectral qualitative identification method combined with analytic hierarchy process, and verifies the feasibility of the method by taking four common food additives (xylitol, L-alanine, sorbic acid, and benzoic acid) and two illegal additives (melamine, and Sudan Red No. I) as the objects of study. Firstly, the collected terahertz time-domain spectral data were pre-processed and transformed into a data set consisting of peaks, peak positions, peak numbers and overall trends; then, the data were divided into comparison and test sets, and a qualitative additive identification model incorporating analytic hierarchy process was constructed and parameter optimisation was performed. The results showed that the qualitative identification accuracies of additives based on single factors, i.e., overall trend, peak value, peak position, and peak number, were 80.23%, 70.93%, 67.44%, and 40.70%, respectively, whereas the identification accuracy of the analytic hierarchy process qualitative identification method based on multi-factors could be improved to 92.44%. In addition, the fuzzy characterisation of the absorption spectrum data was binarised in the data pre-processing stage and used as the base data for the overall trend, and the recognition accuracy was improved to 94.19% by combining the fuzzy characterisation method of such data with the hierarchical analysis qualitative recognition model. The results show that it is feasible to use terahertz technology to identify different varieties of additives, and this paper constructs a hierarchical analytical qualitative model with better effect, which provides a new means for food additives detection, and the method is simple in steps, with a small demand for samples, which is suitable for the rapid detection of small samples.
You may also start an advanced similarity search for this article.