近红外光谱法定性检测餐桌废弃油脂的掺伪

Qualitative identification of illegal cooking oil using near infrared spectroscopy

摘要: 目前餐桌废弃油脂的违法掺伪行为时有发生。探索快速而有效的分析手段,保障百姓餐桌安全是当前的重要任务。本实验采用近红外光谱分析法对餐桌废弃油脂掺伪食用油进行定性鉴别。采用计量学软件,运用距离判别法和BP神经网络算法,对经过标准化与主成分分析处理过的油脂光谱数据建立模式识别模型。距离判别法的判别准确率为89.71%,BP神经网络准确率最高达到97.06%。预测结果说明,分析油脂的近红外光谱信息可以有效鉴别回收油的掺伪行为。

Abstract: Currently, illegal cooking oil cheating occurs constantly. It's important to search a analysis method which is fast and effective for people's table safety. In this study, near infrared spectroscopy was used to discriminate illegal cooking oil cheating. Datas of oil, which was processed with Standardization and PCA, was used for building distance analysis model and BP neural network model by metrology software. The accuracy of distance analysis was 89.71%, and BP neural network was up to 97.06%. In a conclusion, the cheating could be effectively discriminated by analyzing NIRS information.