FT-MIR法鉴别植物油和精炼潲水油的影响因素分析

Influence Factors for Identification of Refining Hogwash Oils and Edible Oils by FT-MIR

采用傅里叶变换中红外光谱（FT-MIR）和偏最小二乘判别法分析技术,对正常食用植物油（EOs）和精炼潲水油（RHOs）的鉴别分析模型的影响因素进行系统分析研究。通过FT-MIR采集4种EOs（菜籽油、大豆油、花生油和玉米油）和RHOs的光谱数据,研究光谱波段范围、光谱数据预处理方法和主成分数选择等因素对判别分析模型的影响。结果表明,筛选出建模光谱波段为4000～450cm-1,光谱数据预处理方法为二阶求导（Savitzky-Golay,5点）,最适主成分数为7;在优化条件下,所建立的基于FT-MIR的偏最小二乘判别法效果最好。

Fourier transform-infrared spectra（FT-MIR） and partial least squares-discriminant analysis technique was applied to compare the influence factors of identification of refining hogwash oils（RHOs） and edible oils（EOs）. FT-MIR was applied to rapidly distinguish between four different edible vegetable oils（rapeseed oil, soybean oil, peanut oil and corn oil） and RHO. The influences of spectra waveband, pretreatment method and principal components on the identification of EOs and RHOs were investigated. The results showed that spectra waveband is 4000 ～450cm-1, pretreatment method second derivative（Savitzky-Golay, 5 point） transformation, and principal components 7. The best FT-MIR identification model was built under the above optimum conditions.