基于可见/近红外光谱技术的牛肉含水率无损检测

Nondestructive determination of water content in beef using visible/near-infrared spectroscopy

生鲜牛肉的含水率对其牛肉的加工、储藏、贸易与食用质量有重要影响,为了提高牛肉的经济价值和食用品质,需要研究牛肉含水率的无损检测技术。以取自不同超市的内蒙小黄牛和鲁西黄牛背最长肌为研究对象,有效样本86个,其中,75%的样本作为校正集,25%的样本作为验证集。采集牛肉新鲜切口处400～1170nm波长范围内的漫反射光谱,用国标方法测定牛肉含水率。经过多元散射校正(multiplicative scatter correction,MSC)、变量标准化(standard normalized variate,SNV)和直接正交信号校正(direct orthogonal signal correction,DOSC)等方法预处理,在400～1170nm范围内分别建立多元线性回归(multiple linear regression,MLR)模型、主成分回归(principal component Regression,PCR)模型和偏最小二乘回归(partial least squares regression,PLSR)模型。结果表明使用MSC预处理方法建立的模型预测效果最佳,其中用PLSR建模结果最好,校正集的相关系数和校正标准差分别是0.92和0.0069,验证集的相关系数和验证标准差分别是0.92和0.0047,外部验证的相关系数和验证标准差分别是0.85和0.0054。结果表明,可见/近红外光谱结合MSC预处理方法建立的PLSR模型,可以对牛肉含水率进行准确的快速无损评价,为生鲜牛肉含水率快速无损检测技术的应用提供理论参考。

The water content of fresh beef has an important influence on the processing, storage, trade and quality of beef. In order to improve the economic value of beef and eating quality, we should research nondestructive testing technology on water content in beef. A laboratory visible/near-infrared spectroscopy system using visible/near-infrared spectroscopy was build to collect 86 beef samples’reflectance spectra in a rang of 400-1170 nm. The samples are from Inner Mongolia cattle’s and Luxi cattle’s longissimus dorsi in different carcasses for the study, 75% of the samples are used as a calibration set, 25% of the samples are used as a validation set. The diffuse reflectance spectra in the fresh cut of beef were collected, and the water contents of the samples were measured with the national standard. The diffuse reflectance spectra of samples were performed with different pretreatments, such as multiplicative scatter correction （MSC）, standard normalized variate （SNV） and direct orthogonal signal correction （DOSC）. The prediction model of multiple linear regression （MLR）, principal component regression （PCR） and partial least squares regression （PLSR） were constructed for prediction of water content in beef with full-spectrum. Correlation coefficient and standard error between prediction water content and real water content of the samples are taken as evaluation criterions for the prediction modal. In general, the higher correlation coefficient of calibration set with validation set and lower standard error of calibration set with validation set mean higher precision of prediction model. Result shows that multiplicative scatter correction is the best pretreatment, and the performance of models established with PLSR is better than others, its correlation coefficient and standard deviation are 0.92 and 0.0047, respectively. The correlation coefficient and standard deviation of external validation set in PLSR model is 0.85 and 0.0054, respectively. Direct orthogonal signal correction combining with principal component regression and partial least squares regression has a high correlation coefficient in calibration set, but a low correlation coefficient in validation set, because of overfitting. This study demonstrated that the PLSR model built by using visible/near-infrared spectroscopy with multiplicative scatter correction pretreatment can nondestructively and rapidly determine the water content in beef. This research can provide a basis for further developing device of nondestructive and rapid determination of water content in beef.