论文部分内容阅读
目的建立蜂蜜样品真伪鉴别的近红外光谱快速检测方法,为今后蜂蜜检验工作提供可靠参考依据。方法采用积分球透反射模式采集样品近红外光谱数据,以Savitzky-Golay 1阶微分方法对原始光谱进行预处理,以随机森林方法建立光谱数据与蜂蜜真伪的定性判别模型。结果所建立的判别模型中训练样本判别正确率为100%,测试样本判别正确率为95%。结论近红外透反射光谱技术应用于蜂蜜真伪鉴别的可行性,同其他分析方法相比具有操作简单、速度快、效率高、无污染、费用低、无需复杂前处理等优点。
Objective To establish a rapid and near-infrared spectroscopy detection method for the authenticity of honey samples and provide a reliable reference for future honey testing. Methods The near-infrared spectral data of the sample were collected by using the integral sphere transflective mode, and the original spectra were preprocessed by the Savitzky-Golay 1-order differential method. A qualitative discriminant model of spectral data and honey authenticity was established by random forest method. Results The discrimination accuracy of the training samples was 100% and that of the test samples was 95%. Conclusion The feasibility of near-infrared trans-reflection spectroscopy applied to the authenticity identification of honey is more simple, faster, more efficient, less pollution, less cost and without complicated pretreatment than other analytical methods.