便携式近红外光谱仪云端智能快速检测花椒饮片

Portable near infrared spectrometer for smart analysis of Zanthoxylum bungeanum Maxim based on cloud technology

目的采用国产的便携式近红外光谱仪及配套的光谱建模系统,针对3个不同产地的药典花椒品种开发定性识别模型并进行验证。方法使用了62个不同产地的花椒样品作为建模集,14个盲样作为验证集,对花椒颗粒进行直接漫反射采样,每个样本采样6次,每次采样取不同的位置取平均作为建模用光谱数据,采用LDA定性识别算法开发花椒产地识别模型。结果模型交叉验证准确率95%以上,盲样验证准确率达到100%。结论对国产便携式近红外光谱仪和配套的光谱建模系统进行了验证,实验效果良好,可以准确识别不同产地的花椒颗粒,在中药材快检领域具备巨大的应用前景。

Objective To qualitatively verify 3 types of Zanthoxylum bungeanum Maxim from different production areas by portable near-infrared spectrometer and spectrum modeling system(made in China). Methods Totally 62 types of Zanthoxylum bungeanum Maxim from different production areas were used as calibration data, and 14 types of Zanthoxylum bungeanum Maxim as the validation data. Reflection method was used to directly sample with the spectrum data, and each was sampled 6 times with each time on different positions and the average data were used as raw spectrum for calibration. LDA method was used to develop qualitative analysis model for different production areas. Results A 95% accuracy of cross validation inside the model was obtained with a 100% accuracy in unknown sample testing. Conclusion Portable near infrared spectrometer and spectrum modeling system made in China can precisely identify Zanthoxylum bungeanum Maxim from different production areas, demonstrating a remarkable prospect in the fast inspection of Chinese medicine.