论文部分内容阅读
目的实现生鲜牛肉新鲜度等级的无损快速判别。方法用可见/近红外光谱检测系统,获取储存1～18d的36块牛肉样品的400～1600nm范围的光谱信息,以挥发性盐基氮理化值为分类依据。用多元散射校正(MSC)、变量标准化(SNV)、SG平滑预处理方法处理光谱数据,分别建立牛肉新鲜度的支持向量机分类模型。结果 MSC+SG预处理后所建立的分类模型预测能力最好,训练集和测试集的回判识别率和预测识别率分别为96.30%、100%,验证集的识别率为88.89%。结论可见/近红外光谱结合支持向量机,对牛肉新鲜度进行无损快速判别是可行的。
Objective To achieve the freshness grade of fresh beef nondestructive rapid discrimination. Methods Spectral information in the range of 400 ~ 1600 nm of 36 beef samples stored for 1 ~ 18 days was obtained by visible / near infrared spectroscopy system. The spectral data were processed by multivariate scatter correction (MSC), variable normalization (SNV) and SG smoothing preprocessing to establish the SVM classification model of beef freshness. Results The classification model established by MSC + SG pretreatment had the best predictive ability. The recognition rate of return and prediction of training set and test set were 96.30% and 100% respectively, and the recognition rate of validation set was 88.89%. Conclusion Visible / near infrared spectroscopy combined with support vector machines, nondestructive rapid determination of freshness of beef is feasible.