To assess the tenderness of nondestructive fresh beef rapidly, 40 beef samples’ reflectance spectra in 400~1650 nm were collected by a laboratory visible and near-infrared spectroscopy system. The Warner-Bratzlar shear force( WBSF) values were measured and taken as the basic of classification. The multiplicative scatter cor-rection ( MSC) , savitzky-golay ( SG) smoothing method were taken as the pretreatment method for reflectance spectra processing respectively. The classifier models of linear discriminant analysis( LDA) and support vector machine( SVM) were built for identification of beef tenderness. The result shows that the MSC combined with SG smoothing was the best pretreatment and the predictive ability of LDA model was the best. The identification rates of training set and prediction set were 96. 7% and 100% respectively. This suggests that combination of spectro-scopic technique and LDA can be used to assess the tenderness of nondestructive fresh beef rapidly.%为实现生鲜牛肉嫩度等级的无损快速判别，用可见近红外光谱检测实验系统，获取不同嫩度的40块牛肉样品的400~1650 nm范围的光谱信息，以肉类嫩度测量仪测量牛肉剪切力值为分类依据.用多元散射校正( MSC)、SG平滑处理方法处理光谱数据，分别建立牛肉嫩度的线性判别分类模型和支持向量机分类模型.结果显示，用MSC+SG预处理后所建立的线性判别分类模型预测能力最好，训练集和测试的回判识别率和预测识别率分别为96.7%、100%.结果表明，用光谱技术结合线性判别分析，可以对牛肉嫩度进行无损快速判别.
展开▼
