作　　者： (孙梅）; (王睿）; (陈兴海）; (黄宇）; (刘业林）;

机构地区： 北京工商大学计算机与信息工程学院,北京100048

出　　处： 《中国现代应用药学》 2017年第8期1114-1117,共4页

摘　　要： 目的以冬虫夏草粉末为研究对象,利用高光谱成像技术建立对冬虫夏草粉末的真假鉴别及含量判断的无损检测模型。方法将真伪样品粉末分别按9︰1、8︰2、7︰3、6︰4、5︰5等不同比例进行混合,通过光谱范围为940~2 500 nm的高光谱成像仪分别获取真伪样品、不同比例真伪混合样品的高光谱影像数据,经过主成分分析(PCA)初步判别真伪样品的差异性,再结合偏最小二乘法(PLS)对真伪粉末样品的含量进行分析判断。结果不同比例真伪混合物中冬虫夏草粉末的识别度为97.0%~98.78%,伪冬虫夏草粉末的识别度为83.10%~99.3%。结论基于高光成像技术可以实现对冬虫夏草粉的真假辨别,并可准确地判别出冬虫夏草粉末的有效含量。 OBJECTIVE To identify true or false of cordyceps powder and judge it＇s content use the hyperspectral imaging technology. METHODS The true and false samples powder were mixed at different ratios of 9︰1, 8︰2, 7︰3, 6︰4, 5︰5, respectively. The hyperspectral image data of samples and different proportions of true and false mixed samples were obtained by the spectral range of 940-2 500 nm hyperspectral imager. Through principal component analysis（PCA） to determine the difference between true and false samples, and combined with partial least squares（PLS） method to analyze the content of true and false powder samples. RESULTS In different proportions of true and false mixtures, cordyceps powder identification degree was 97.0%-98.78%, and false cordyceps powder identification degree was 83.10%-99.3%. CONCLUSION The research results show it is convenient to identify the true and false of cordyceps powder by the hyperspectral imaging technology, and it is suitable for quality evaluation of traditional cordyceps powder.

关 键 词： 高光谱成像技术 冬虫夏草粉末 真假鉴别 含量判断