首页|基于近红外光谱技术和LSPTSVM模型的镉污染贻贝检测研究

基于近红外光谱技术和LSPTSVM模型的镉污染贻贝检测研究

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
贝类重金属污染已成为亟待解决的海洋食品安全问题,镉是重要的污染源之一.食用被重金属镉污染的翡翠贻贝会严重危害身体健康.研究了基于近红外反射光谱的镉污染贻贝无损、快速检测方法.通过采集正常贻贝和镉污染贻贝的950～1700 nm光谱数据,构建基于最小二乘投影孪生支持向量机(least squares projection twin support vector machine,LSPTSVM)检测模型,优化模型参数和正交投影轴数量获得最佳检测性能.提出的LSPTSVM模型检测镉污染贻贝准确率达到99.50％,优于其他孪生支持向量机模型.LSPTSVM模型适用于小样本数据集.针对难以获得大量镉污染训练样本情况,LSPTSVM模型较其他模型具有更好的稳健性.结果表明:近红外光谱结合LSPTSVM模型可实现镉污染贻贝检测,为贝类的品质评价和安全检测提供一种新的方法.
DETECTION OF MUSSELS CONTAMINATED WITH CADMIUM BASED ON NEAR-INFRARED SPECTROSCOPY AND LSPTSVM
Heavy metal contamination of shellfish has become an urgent problem of marine food safety,among which cadmium is one of the important contamination sources.The consumption of mussels contaminated with heavy metal cadmium is a serious health hazard.A non-destructive and rapid detection method for mussels contaminated with cadmium based on near-infrared reflectance spectroscopy was researched in this study.By collecting spectral data of normal and cadmium-contaminated mussels in the range of 950～1700 nm,a detection model based on the least squares projection twin support vector machine(LSPTSVM)was constructed.The parameters of the model and the number of orthogonal projection axes were optimized to obtain the best detection performance.The accuracy of the proposed LSPTSVM model achieved 99.50％for detecting cadmium-contaminated mussels,which was superior to other twin support vector machine models.The LSPTSVM model was suitable for the datasets with small samples.In the case that it was difficult to obtain many cadmium-contaminated training samples,the LSPTSVM model had better robustness than other models.The results showed that near-infrared spectroscopy combined with the LSPTSVM model can realize the detection of cadmium-contaminated mussels,which provides a new method for quality evaluation and safety detection of shellfish.

near-infrared spectroscopycadmium contaminationleast squares projection twin support vector machinemusselsheavy metal detection

曾绍庚、刘瑶、刘忠艳

展开 >

岭南师范学院计算机与智能教育学院,广东湛江 524048

岭南师范学院电子与电气工程学院,广东湛江 524048

近红外光谱 镉污染 最小二乘投影孪生支持向量机 贻贝 重金属检测

国家自然科学基金青年科学基金广东省自然科学基金面上项目广东省教育厅普通高等学校新冠肺炎疫情防控专项研究项目广东省科技创新战略专项(大学生科技创新培育)立项项目

620051092020A5150113682020KZDZX1143pdjh2024a240

2024

10.13205/j.hjgc.202401030

环境工程
中冶建筑研究总院有限公司,中国环境科学学会环境工程分会

环境工程

CSTPCD
影响因子：0.958
ISSN：1000-8942
年,卷(期)：2024.42(1)
  • 41