傅立叶近红外光谱法测定桃果中果胶含量

Determination of pectin in peach fruit by Fourier transform near infrared spectroscopy

中文关键词:  桃  果胶  傅立叶近红外光谱技术  光谱预处理  建模

英文关键词:peach  pectin  Fourier transform near infrared spectroscopy  spectral pretreatment  modeling

中文摘要:

      目的 建立傅立叶近红外光谱法(Fourier transform near infrared spectroscopy, FT-NIR)检测桃果中果胶含量的分析方法。方法 以2个桃品种为材料, 采集近红外光谱样品, 探讨光谱预处理对模型的影响。采用偏最小二乘法(partial least squares, PLS)以及主成分回归(principal component regression, PCR)进行建模, 建模相关系数(rc)、建模均方偏差(root mean square error of calibration, RMSEC)、预测相关系数(rp)、预测均方偏差(root mean square error for prediction, RMSEP)作为模型的评价标准。结果 2个品种的近红外光谱图和果胶含量无明显差异(P＞0.05), 采用标准正态变量变换(standard normal variate, SNV)和多元散射校正(multiplicative signal correction, MSC)对原始光谱的光程进行选择, 对建模结果的影响基本一致。综合得出最佳模型是利用PLS方法建模并采用MSC/SNV结合一阶导数和Savitzky-Golay (S-G)平滑对近红外光谱图进行预处理, 评价参数分别为rc=0.7795、rp=0.7545、RMSEC=0.0933、RMSEP=0.0534和rc=0.7800、rp=0.7530、RMSEC=0.0932、RMSEP=0.0534。结论 该方法准确、可靠, 适用于桃果中果胶含量的快速检测。

英文摘要:

      Objective To establish a method for the determination of pectin content in peach fruit by Fourier transform near infrared spectroscopy(FT-NIR). Methods Using 2 peach varieties as materials, samples of near-infrared spectroscopy were collected to discuss the influence of spectral pretreatment on the model. Modeling was conducted by partial least squares (PLS) and principal component regression (PCR) methods. Modeled correlation coefficient (rc), modeling mean square deviation (RMSEC), predicted correlation coefficient (rp), and predicted mean square deviation (RMSEP) were the evaluation criteria for the model. Results There was no significant difference between the NIR spectra and pectin content of 2 varieties (P>0.05). The standard normal variable transformation (SNV) and multiple scattering correction (MSC) were used to select the optical path of the original spectrum. The results of the model were basically consistent. It was concluded that the best model was to use the PLS method to model and use MSC/SNV combined with the first derivative and Savitzky-Golay (SG) smoothing to preprocess the near infrared spectrum. The evaluation parameters were rc=0.7795, rp=0.7545, RMSEC=0.0933, RMSEP=0.0534 and rc=0.7800, rp=0.7530, RMSEC=0.0932, RMSEP=0.0534. Conclusion This method is accurate, reliable and suitable for the rapid determination of pectin content in peach fruit.