Study On Quality Evaluation For Honey By Near Infrared Spectroscopy An D Mid-Infrared Spectroscop

In recent years, fourier transform infrared spectroscopy is widely used in agriculture, food industry, chemistry, pharmaceutical industry, textile industry. Fourier transform near-infrared reflectance (FT-NIR) spectroscopy with chemometrics was used in the research to determine the contents of fructose and glucose in honey. Fourier transform near-infrared reflectance (FT-NIR) and fourier infrared mid-infrared reflectance (FT-MIR) spectroscopy with chemometrics were used for floral classification of honey samples respectively. FT-NIR and FT-MIR spectroscopy was used to determine high fructose corn syrup adulterated in honey. Partial least square (PLS) regression and several discriminate models were established.Fructose and glucose of honey were determined by PLS regression using near infrared spectroscopy with a fiber optic probe. For PLS regression of fructose, a correlation of0.9987was obtained between predicted and reference values (by HPLC) with root mean square error of cross-validation (RMSECV) of0.246. For PLS regression of glucose, a correlation of0.9716was obtained between predicted and reference values (by HPLC) with root mean square error of cross-validation of0.544. The data of validation were analyzed by t-test which demonstrated good reliability and accuracy of the models.FT-NIR and FT-MIR spectroscopy with chemometrics were used for floral classification of honey samples respectively. Discriminant function analysis (DFA) and Bayesian function analysis was conducted on four honey varieties according to nectar sources (vitex, acacia, peach blossom, and jujube). FT-NIR spectrometer equipped with a fiber optic probe with Bayesian discriminant analysis and DFA were applied to floral classification of honey samples, which gained total correct classification rates of100%. Meanwhile, Bayesian discriminant analysis was applied to spectral data gained from FT-MIR spectrometer equipped with an attenuated total reflectance accessory. The total correct classification rates of calibration and validation were97.92%and93.75%, respectively. The result indicates that both NIR and MIR spectroscopy with chemometrics can be a useful tool for honey classification according to nectar sources and could be an alternative to traditional method. Furthermore, the classification model with NIR spectroscopy is stronger than that with MIR spectroscopy.FT-NIR and FT-MIR spectroscopy with chemometrics were used to determine high fructose corn syrup adulterated in honey. Discriminate function analysis (DFA) and Bayesian function analysis were conducted to classify adulterated honey and pure honey. Adulterated honey was classified when high fructose corn syrup adulterated was equal or greater than15%. Content of high fructose corn syrup was predicted by PLS regression. A correlation of0.9935was obtained between predicted and reference values with root mean square error of cross-validation of2.56using FT-NIR. A correlation of0.9661was obtained between predicted and reference values with root mean square error of cross-validation of5.4using FT-MIR. Reliability of models was verified by external-validation. The result demonstrated good reliability and accuracy of the models.