论文部分内容阅读
多源光谱分析技术被用于鱼油品牌快速无损鉴别。采用可见光谱分析技术、短波近红外光谱分析技术、长波近红外光谱分析技术、中红外光谱分析技术和核磁共振光谱分析技术采集了7种不同品牌的鱼油的光谱特征,并应用偏最小二乘判别分析法(partial least squares discrimination analysis,PLS-DA)和最小二乘支持向量机(least-squares support vector machine,LS-SVM)建立判别模型并比较判别结果。基于长波近红外光谱的PLS-DA模型和LS-SVM模型取得了最高识别正确率,建模集和预测集识别正确率均达到100%。采用中红外光谱和核磁共振谱分别建立的LS-SVM模型,也可以获得100%的判别正确率。而可见光谱和短波近红外光谱则判别准确率较差。且LS-SVM算法较PLS-DA更加适合用于建立光谱数据和鱼油品牌之间的判别模型。研究结果表面长波近红外光谱技术能够有效判别不同鱼油的品牌,为将来鱼油品质鉴定便携式仪器的开发提供了技术支持和理论依据。
Multi-source spectroscopic techniques are used to quickly and easily identify fish oil brands. The spectral characteristics of seven different brands of fish oil were collected by visible spectrum analysis, shortwave near-infrared spectroscopy, long-wavelength near-infrared spectroscopy, mid-infrared spectroscopy and nuclear magnetic resonance spectroscopy. Partial least square discriminant Partial least squares discrimination analysis (PLS-DA) and least-square support vector machine (LS-SVM) were used to establish the discriminant model and compare the discriminant results. The highest recognition accuracy of PLS-DA model and LS-SVM model based on longwave near-infrared spectroscopy was achieved, and both the model set and the prediction set recognition accuracy reached 100%. Using LS-SVM model established by mid-infrared spectroscopy and nuclear magnetic resonance spectroscopy, the discrimination accuracy of 100% can be obtained. The visible spectrum and shortwave near infrared spectroscopy to determine the accuracy of poor. And the LS-SVM algorithm is more suitable than PLS-DA for establishing the discriminant model between the spectral data and the fish oil brand. The results of the surface long-wave near-infrared spectroscopy technology can effectively distinguish different brands of fish oil, providing technical support and theoretical basis for the future development of fish oil quality identification portable instrument.