食用油脂酸值近红外光谱特征波长优选

以近红外光谱快速检测大豆油脂酸值为目标,研究间隔偏最小二乘(interval partial least square,i PLS)结合遗传算法(genetic algorithm,GA)及连续投影算法(successive projection algorithm,SPA)的特征波长变量优选方法。制备不同酸值的大豆油脂样品100个,并在4 000~12 000 cm-1范围内采集了油样的近红外透射光谱。首先用i PLS法从原始光谱中初步筛选出4 540~5 346 cm-1和6 807~7 004 cm-1组合特征波段,R2和预测均方根误差(root mean square error of prediction,RMSEP)分别为0.978 9和0.064 3;然后分别用GA和SPA从特征光谱区域中筛选出与油脂酸值密切相关的特征波长变量,从GA和SPA 2种选择结果中各选取前6个波长点,以12个特征波长变量建立PLS校正模型,其R2和RMSEP分别为0.985 9和0.045 1。研究表明,在油脂酸值近红外光谱分析中,采用i PLS-GASPA相结合的方法进行特征波长选择能有效去除冗余信息,降低模型复杂度,可为快速无损检测油脂酸值提供重要理论依据。     

近红外光谱结合区间偏最小二乘法应用于花生油酸价的测定

采用近红外光谱法结合不同区间偏最小二乘波长筛选法建立花生油酸价的定量分析模型。采用酸碱滴定法测定花生油样本的酸价同时采集近红外光谱数据;采用区间偏最小二乘法(iPLS)、向后区间偏最小二乘法(BiPLS)、移动窗口偏最小二乘法(mwPLS)优选光谱特征区间;采用偏最小二乘法(PLS)对优选出来的谱段建立酸价的定量模型。结果表明,采用mwPLS选择的谱段建立的模型预测效果最佳,RMSECV和RMSEP分别为0.247 76和0.131 5,校正相关系数和预测相关系数分别为0.993 2和0.996 9。因此,近红外光谱结合移动窗口偏最小二乘法可以快速准确测定花生油的酸价。     

Real-time pixel based early apple bruise detection using short wave infrared hyperspectral imaging in combination with calibration and glare correction techniques

High speed data processing for online food quality inspection using hyperspectral imaging (HSI) is challenging as over hundred spectral images have to be analyzed simultaneously. In this study, a real-time pixel based early apple bruise detection system based on HSI in the shortwave infrared (SWIR) range has been developed. This systems consists of a novel, homogeneous SWIR illumination unit and a line scan camera. The system performance was tested on Jonagold apples bruised less than two hours before scanning. Partial least squares-discriminant analysis was used to discriminate bruised pixel spectra from sound pixel spectra. As the glossiness of many fruit and vegetables limits the accuracy in the detection of defects, several reflectance calibrations and pre-processing techniques were compared for glare correction and maximizing the signal to noise ratio. With the best combination of first derivative and mean centering, followed by image post-processing, this system was able to detect fresh bruises in thirty apples with 98% accuracy at the pixel level with a processing time per apple below 200 ms.     

Multivariate near-infrared and Raman spectroscopic quantifications of the crystallinity of lactose in whey permeate powder

The use of near-infrared (NIR) and Fourier transform Raman spectroscopy for quantification of crystalline lactose content in whey permeate powder was investigated using chemometric methods. Sample sets consisting of binary mixtures of crystalline (50.0–98.0%) and amorphous lactose and process whey permeate samples with different amounts of crystalline lactose (75.0–95.5%) added were analyzed. The best results for quantification of crystallinity were obtained by partial least squares (PLS) regression on NIR data in five selected intervals in the range 1100–2498 nm. Data analysis on the total sample set of 35 samples yielded a prediction error (root mean square error of cross validation) of 0.627%. The corresponding result for Raman spectroscopy in the range 3500–100 cm⁻¹ was 1.62%. Interval-PLS regression was used for the selection of relevant spectral intervals as well as for improving the spectral interpretation. Alternating regression was used to show that the amorphous lactose preparation contained only a negligible amount of crystalline lactose.     

基于iPLS的油脂过氧化值近红外光谱特征波段选择

在大豆油脂过氧化值近红外光谱分析中,利用间隔偏最小二乘法(interval partial least square,iPLS)实现油脂光谱特征波段选择。分别将全谱波段以10个数据点间隔和20个数据点间隔分成若干个小波段,然后对全谱和每个小波段分别用PLS回归建模,用预测残差平方和(predicted residual sum of squares,PRESS)对模型进行评价。结果表明:经过特征波段选择后,50个波长点模型的决定系数、预测误差均方根、相对误差均值分别为0.9791、0.0513和2.12%,有效地减少建模的变量数,预测精度得到提高。     

基于iPLS的玉米近红外光谱数据处理方法研究

通过对选取的离散小波变换后的小波细节系数的重构光谱进行iPLS建模,从中获得了较好的预测结果。实验表明,合理选择小波细节系数可减少非主要信息的干扰,并以此来提高成分分析精度的方法是切实可行的,采用iPLS建立的玉米蛋白质模型,不仅能有效地减少建模所用的变量数,缩短运算时间,而且预测能力和精度均得到提高,在实际应用中有一定的参考价值。     

基于iPLS原理最优化信息区间的桃糖度组合权重PLS模型研究

采用反向区间偏最小二乘法和组合区间偏最小二乘法优化桃糖度可见/近红外光谱的信息区间组合,在选择的信息区间的基础上建立了一种线性组合权重PLS模型.对近红外光谱进行二阶导数处理、卷积平滑校正后,发现在区间分割数为15时筛选结果最优,BiPLS所选择的信息区间为742~770nm和 862~920nm,SiPLS所选择的信息区间为742~770nm、832~860nm和892~920nm.直接组合信息区间BiPLS和SiPLS模型的RMSEP值分别为0.386和0.308,线性组合权重PLS模型的RMSEP值分别为0.351和0.364.结果说明在近红外定量分析中线性组合权重模型的建立克服了复杂样品各信息区间对PLS建模贡献率不一样的问题.     

基于iPLS原理最优化信息区间的桃糖度组合权重PLS模型研究

采用反向区间偏最小二乘法和组合区间偏最小二乘法优化桃糖度可见/近红外光谱的信息区间组合,在选择的信息区间基础上提出了建立一种线性组合权重PLS模型的方法。对近红外光谱进行二阶导数处理、卷积平滑校正后,在区间分割数为15时筛选结果最优,BiPLS所选择的信息区间为742~770nm和 862~920nm,SiPLS所选择的信息区间为742~770nm、832~860nm和892~920nm;直接组合信息区间BiPLS和SiPLS模型的RMSEP值分别为0.386和0.308;线性组合权重PLS模型的RMSEP值分别为0.351和0.364。结果说明在近红外定量分析中,建立线性组合权重模型具有克服复杂样品各信息区间对PLS建模贡献率不一样的问题。     

FTIR技术结合区间偏最小二乘法快速测定油脂中反式脂肪酸

研究了傅里叶红外光谱技术结合区间偏最小二乘法(iPLS)快速分析食用油中低含量(0.1%~5%)反式脂肪酸的分析方法。通过系统地比较衰减全反射红外光谱法(ATR-FTIF)及衰减透射红外光谱法(TR-FTIR)光谱的模型效果,优化建模区间。研究结果表明,ATR-FTIR、TR-FTIR-PLS回归模型均能有效测定油脂中低浓度反式脂肪酸的含量,但TR-FTIR法灵敏度优于ATR-FTIR法。iPLS区间选择结果显示,以1 000~940 cm~(-1)波段透射光谱建模,相关系数R~2为0.998 8,标准集的RMSEC 0.016 6,验证集RMSEP为0.008 75,预测相对标准偏差2.92%,预测值与实际值高度相关,Y_(预测)=1.00X_(实际)-0.003 44,R~2=0.998 7。12组外部验证试验相对标准偏差为4.80%,说明预测精确度较高、模型稳定性好,有潜力替代传统气相色谱法用于油脂中低含量反式脂肪酸快速定量测定。