试验探讨了近红外反射光谱测定大豆制品中寡糖(蔗糖、 棉籽糖及水苏糖)含量的可行性。 2012年—2014年, 从国内20个代表性大豆制品加工厂收集并筛选了160个大豆制品(包括去皮豆粕、 膨化豆粕、 发酵豆粕及膨化大豆各40个), 同时为使样品集中的寡糖含量均匀分布, 采用发酵豆粕和普通豆粕混合的方法, 配制人工混合豆粕样品40个, 使用偏最小二乘法, 采用不同导数或去散射方法对光谱进行预处理并建立近红外定标模型。 在光谱预处理中, 针对不同寡糖, 导数处理、 多元散射校正及矢量归一化方法得到良好的预测效果。 所建立的近红外定标模型的蔗糖、 棉籽糖和水苏糖的定标决定系数(R2cal)分别为0.99, 0.95和0.98, 交互验证决定系数(R2cv)分别为0.98, 0.94和0.97, 交互验证RPDcv值分别为7.24, 4.13和5.98, 表明适用于大豆制品(除发酵豆粕外)中寡糖的日常检测。

The present study investigated the possibility of using near-infrared reflectance spectroscopy (NIRS) for quantitative determination of oligosaccharides, including sucrose, raffinose and stachyose in soybean products. One hundred and sixty samples (de-hulled soybean meal, extruded soybean meal, extruded full-fat soybean and fermented soybean meal) were collected and screened, and furthermore forty artificial soybean meal samples were prepared. Calibrations were performed by partial least squared algorithm with different derivatives, multiplicative scatter correction or normalization vector spectral pretreatment. The coefficients of determination for calibration (R2cal), coefficients of determination for cross validation (R2cv) and RPDcv in models were 0.99, 0.95, 0.98; 0.98, 0.94, 0.97; 7.24, 4.13, 5.98; respectively. The results supposed that NIRS can be a quantitative determination method of oligosaccharides in soybean products except fermented soybean meal.