摘要
提出一种基于近红外光谱技术和聚类分析法对不同种类的淀粉进行无损快速鉴别的方法,研究了近红外光谱结合聚类分析法快速鉴别淀粉种类的可行性.采集样品在12 500~4 000cm-1范围内的近红外漫反射光谱,利用光谱模式识别法中的聚类分析法对图谱进行定性分类鉴别.光谱经一阶导数预处理后,采用聚类分析法可准确无误地将样品分为2类,分类准确率达到100%.同时采用预测集的样品验证所建模型的准确性,经实验得出预测准确率也是100%.实验结果表明,近红外漫反射光谱结合聚类分析法可以为快速无损鉴别淀粉种类提供一种准确可靠的方法.
A rapid and undamaged method to identify different kinds of starch based on the near infrared spectroscopy and clustering analysis was put forward and the feasibility was analyzed in this paper.The wavelength range is between 12500cm-1 and 4000cm-1.The spectra of starch were identified qualitatively according to clustering analysis which is a kind of spectra pattern recognition.The result showed that the starch samples could be identified accurately into two kinds with the first-derivative preprocessing spectra by clustering analysis.The classification exactness rate of 100% was achieved.At the same time,the accuracy of the model was verified using the testing set samples and the exactness rate was nearly 100%.The near infrared diffuse reflectance spectroscopy and clustering analysis could be a rapid and undamaged approach to discriminate the species of the starch with great accuracy and reliability.
出处
《北京工商大学学报（自然科学版）》
CAS
2010年第6期53-56,共4页
Journal of Beijing Technology and Business University:Natural Science Edition
基金
北京市自然科学基金资助项目(4073031)
