天津农学院学报 ›› 2025, Vol. 32 ›› Issue (1): 77-81.doi: 10.19640/j.cnki.jtau.2025.01.014

• • 上一篇    下一篇

基于近红外光谱技术的鸡饲料中微塑料检测方法研究

刘一诺, 霍正婷, 刘菊香, 董桂梅, 杨延荣, 杨仁杰通信作者   

  1. 天津农学院 工程技术学院,天津 300392
  • 收稿日期:2024-05-06 发布日期:2025-03-04
  • 通讯作者: 杨仁杰(1978—),男,教授,博士,主要从事光谱检测技术与应用方面的研究。E-mail:rjyang1978@163.com。
  • 作者简介:刘一诺(2002—),女,硕士在读,主要从事光谱检测方面的研究。E-mail:2856493715@qq.com。
  • 基金资助:
    国家自然科学基金(41771357,21607114,81471698)

Research on the detection method of microplastics in chicken feed based on near-infrared spectroscopy

Liu Yinuo, Huo Zhengting, Liu Juxiang, Dong Guimei, Yang Yanrong, Yang RenjieCorresponding Author   

  1. College of Engineering and Technology, Tianjin Agricultural University, Tianjin 300392, China
  • Received:2024-05-06 Published:2025-03-04

摘要: 提出并建立了一种基于近红外光谱技术鸡饲料中微塑料的快速检测方法。分别准备80个鸡饲料样品和80个掺入不同质量百分比浓度(0.01%~0.8%)聚丙烯的鸡饲料样品,采用Perkin Elmer公司的Spectrum GX傅里叶变换近红外光谱仪采集所有样品的近红外漫反射光谱。在研究未掺入和掺入聚丙烯鸡饲料近红外光谱特征的基础上,分别建立了定性定量分析掺聚丙烯鸡饲料的数学模型。对于定性模型,校正集和预测集样品的判别正确率分别为95.2%和92.5%。对于定量模型,其校正相关系数R和校正均方根误差分别为0.98和8.83×10-4,预测相关系数R和预测均方根误差分别为0.97和1.77×10-3。结果表明,基于近红外光谱技术快速检测鸡饲料中的微塑料不仅可行,而且取得较好的分析结果。

关键词: 饲料, 近红外光谱, 微塑料, 定性定量分析, 偏最小二乘

Abstract: A rapid detection method for microplastics was proposed and established in chicken feed based on near-infrared spectroscopy. 80 chicken feed samples and 80 chicken feed samples adulterated with polypropylene at different mass percentage concentrations(0.01%-0.8%)were prepared respectively. The near-infrared spectra of all samples were collected by Perkin Elmer's Spectrum GX Fourier transform NIR spectrometer. On the basis of studying the near-infrared spectral characteristics of chicken feed without and with added polypropylene, the qualitative and quantitative analysis models for chicken feed adulterated with polypropylene were established. For qualitative model, the discrimination accuracies of samples in calibration set and prediction set were 95.2% and 92.5%, respectively. For quantitative model, the calibrated correlation coefficient R and the calibrated root mean square error were 0.98 and 8.83×10-4 respectively, and the predicted correlation coefficient R and the predicted root mean square error were 0.97 and 1.70×10-3 respectively. The results indicated that rapid detection of microplastics in chicken feed based on near-infrared spectroscopy is not only feasible, but also has achieved good analytical results.

Key words: chicken feed, near-infrared spectroscopy, microplastics, qualitative and quantitative analysis, partial least squares

中图分类号: