基于近红外光谱仪与模式识别方法的不同年份陈皮无损鉴别研究

Nondestructive Identification of Different-Age Citri Reticulatae Pericarpium Using Near-Infrared Spectrometer and Pattern Recognition Methods

该研究利用光栅型便携式近红外光谱仪采集不同陈化年份陈皮内囊和外壁的近红外光谱,利用光谱预处理方法结合不同模式识别方法构建不同年份陈皮的鉴别模型。结果表明:预处理方法可以有效消除光谱中存在的多种干扰;主成分分析方法无法实现不同年份陈皮的准确鉴别分析,外壁和内囊数据最优鉴别率仅为35%和44%;采用软独立模式分类法可以得到更加准确的鉴别结果,外壁和内囊数据最优鉴别率分别为94%和96%;Fisher线性判别分析方法结果最优,外壁最优鉴别率为98%,而采用内囊数据结合原始光谱便可实现不同年份陈皮100%鉴别分析。以上结果表明,采用光栅型便携式近红外光谱仪结合化学计量学方法可以实现不同年份陈皮的无损鉴别分析。

The spectra of the inner capsule and outer skin of different-age Citri Reticulatae Pericarpium were obtained using a grating portable near-infrared spectrometer.The identification models of the different-age Citri Reticulatae Pericarpium were established using spectral pretreatment methods combined with different pattern recognition methods.The preprocessing method can effectively eliminate the interferences in the spectra.Accurate identification performance of different-age Citri Reticulatae Pericarpium could not be achieved using only the principal component analysis method.The best identification accuracy of 35%was obtained using the outer skin data,while the best identification accuracy of 44%was obtained using the inner capsule data.More accurate identification results were obtained through soft independent modeling of class analogy.The best identification accuracy of 94%was obtained using the outer skin data,whereas the best identification accuracy of 96%was obtained using the inner capsule data.The Fisher linear discriminant analysis results were the best.Identification accuracy of 98%was obtained using the outer skin data,whereas identification accuracy of 100%was obtained using the raw spectra data of the inner capsule.The above results show that the grating portable near-infrared spectrometer combined with the chemometrics method can achieve the nondestructive identification of different-age Citri Reticulatae Pericarpium.