【摘要】 采用正交信号校正法(OSC)对苹果的近红外光谱(1300nm～2100nm)进行预处理,并结合偏最小二乘法(PLS)建立了苹果光谱对糖度的预测模型。应用结果表明,经OSC法预处理后,光谱形状总体上与原始光谱没有差别,但光谱曲线变得更为光滑,排列更为整齐、紧密。这说明正交信号校正法(OSC)滤除了原始光谱中的部分噪声,但又保留了原光谱中的主要信息。苹果光谱对糖度的PLS校正模型采纳的最佳因子数会随着OSC因子的被逐个滤除而逐渐减少,甚至可减少至1(当然模型精度也有变化)。本研究中,校正模型的最佳性能产生于原始光谱被滤除10个OSC因子时,此时其采纳的最佳因子数为2,校正时的相关系数r2和标准偏差SEC分别为0.92644和0.40250,预测时的标准偏差SEP为0.50229。与OSC法处理前的PLS模型相比,其精度虽没有大幅提高,但由于采纳的因子数少,模型变得十分简洁。更多还原

【Abstract】 Orthogonal signal correction (OSC) was used as a method to preprocess the near infrared (NIR) spectra of apples ranging from 1300nm to 2100nm, to estahlish the calibration model of sugar content against apple spectra before and after OSC pretreatment by partial least square (PLS). Compared with those just being centered, apple spectra after OSC followed by centered pretreatment were smoother and in a closer and more orderly array, but their shape showed not much difference. This indicated that the major information in apple spectra could be reserved while part noise was removed by OSC method. The number of optimal factors of PLS model used to predict sugar content against apple spectra would be reduced in accordance with OSC factors reduction, even up to 1 finally (the precision of model also will have a little variation). In this study, the optimum PLS calibration model was obtained when 10 OSC factors were filtered, to have obtained the correlation coefficient (r 2 ) of 0.92644, with the standard error of calibration (SEC) of 0.40250 and the standard error of prediction (SEP) of 0.50229. Although this model could not improve precision to a great extent, but in comparison on with the model before OSC pretreatment, less reduction factors would be necessary and the model would be simpler.更多还原