论文部分内容阅读
利用近红外光谱仪,提取4种不同品种生鲜羊肉样本的光谱数据共235个。分析不同品种样本在400~1000 nm波长范围内光谱值,确定3段特征波段。利用主成分分析结合判别分析建立鉴别模型,达到无损鉴别生鲜羊肉品种的目的。结果显示,在400~430 nm范围内主成分数为7时,校正集回代准确率为75.5%,交叉验证准确率为73.4%,验证集准确率为93.1%。经过一阶导数与标准归一化处理后所建立的模型,当主成分数为23时,400~430 nm波段判别准确率达到93.6%,交叉验证准确率为89.4%。对比其他几种预处理方法和波段的建模结果,选择最优模型。该研究表明利用近红外光谱分析技术可对不同品种羊肉进行快速准确鉴别。
Using near infrared spectroscopy, extraction of four different varieties of fresh lamb samples a total of 235 spectral data. Analysis of different varieties of samples in the 400 ~ 1000 nm wavelength range of spectral values, to determine the three bands. The principal component analysis combined with discriminant analysis to establish identification model to achieve the purpose of non-destructive identification of fresh mutton varieties. The results showed that the accuracy of the calibration set was 75.5%, the accuracy of cross-validation was 73.4%, and the accuracy of validation set was 93.1% when the number of principal components was in the range of 400-430 nm. After the first derivative and standard normalized model were established, when the number of principal components is 23, the accuracy of discriminating between 400-430 nm band was 93.6% and the cross-validation accuracy was 89.4%. Comparing the modeling results of several pretreatment methods and bands, we choose the optimal model. The research shows that the use of near infrared spectroscopy can be quickly and accurately identify different varieties of mutton.