为实现药品的无损快检,设计了基于傅里叶变换红外光谱检测的光学系统,系统工作波段为4000～12000 cm-1.设计合适的光学结构并确定光学元件的特征参数,用Zemax对设计的光学系统进行光线追迹和优化分析.搭建实验平台,采集了5种相似药品的近红外光谱数据,用主成分分析法(PCA)结合误差反向传播人工神经网络算法(BPANN)对数据进行处理,有效区分了5种药品的光谱,识别准确率达到93.333％.实验结果表明:该近红外光谱检测系统适用于药品的无损分析,能有效鉴别相似药品.
To achieve fast nondestructive detection of drugs,the optical system of infrared spectroscopy detection based on Fourier transform is designed.The operating band of system ranges from 4000 cm-1 to 12000 cm-1.The suitable optical structure is designed and characteristic parameters of the optical components are determined.Ray tracing and optimization analysis on optical system designed are carried out with ZEMAX.The experimental platform is set up,and the near-infrared spectral data of five similar drugs are collected.The data are processed by combining principal component analysis (PCA) with error back propagation artificial neural network (BPANN).The spectra of five drugs are effectively identified and the recognition accuracy ratio reaches 93.333％.Experimental results show that the near-infrared spectroscopy detection system is suitable for non-destructive analysis of drugs,and can effectively identify similar drugs.
参考文献
- 下载量()
- 访问量()
- 您的评分：
-
10%
-
20%
-
30%
-
40%
-
50%