基于近红外光谱技术的大米掺伪定量判别

针对现在市场上常见的两种大米掺伪现象,利用近红外光谱技术结合化学计量学方法分别建立了大米中掺入低档米和掺入矿物油的定量分析模型。制配不同掺伪比例的大米样品,采集其近红外光谱,并选用标准正态变量变换、最大最小归一化、平滑和一阶导数4种方法对原始光谱进行预处理,分别结合偏最小二乘法建立PLS定量分析模型。通过对比建模结果选出的最优预处理方法是最大最小归一化,建立的掺低档米模型的校正集和预测集相关系数分别为0.9698和0.9845,均方根误差分别为8.66和6.46;掺矿物油米模型的校正集和预测集相关系数分别为0.9739和0.9888,均方根误差分别为0.106和0.0698。模型的预测精度和稳定性均很好,实现了对两种掺伪大米快速、准确的定量判别,为大米的品质监控提供了一种新的方法思路。

Quantitative determination of adulteration in rice based on near infrared spectroscopy analysis

In view of the two common adulteration phenomena in the rice market, quantitative analysis models were established to detect the rice mixed with low quality rice or mineral oil based on near infrared spectroscopy combined with chemometrics techniques rapidly. The samples of rice adulteration were prepared by mixing high quality rice with low quality rice or mineral oil in different proportions. The near infrared spectra of samples (rice adulteration by low quality rice) were collected, and then, the Maximum-minimum normalization, Standard normal variate, 1st derivative and Smoothing were used to preprocess the spectral data. Partial least squares (PLS) model was developed for predicting the adulteration ratio. The results showed that the model using Maximum-minimum normalization pretreatment method was the best, with the coefficients (R_c, R_p) from calibration set and prediction set were 0.9698 and 0.9845, respectively, and the root mean square error of calibration (RMSEC) and root mean square error of prediction (RMSEP) were 8.66 and 6.46, respectively. In addition, another PLS model based on the rice mixed with mineral oil was established. The R_c and R_p were 0.9739 and 0.9888, respectively, while the RMSEC and RMSEP were 0.106 and 0.0698, respectively. The results indicated that NIR spectroscopy can be applied for a rapid detection of the adulterated rice.