基于LIF对水包油型轻质油乳化液的分类识别  

Classification and Identification of Oil-in-Water Light Oil Emulsions Based on LIF

作　　者：陈晓玉[1] 宁晓东 李心怡 杜雅欣 孔德明[1] CHEN Xiao-yu;NING Xiao-dong;LI Xin-yi;DU Ya-xin;KONG De-ming(Department of Electronic and Communication Engineering,School of Information Science and Engineering,Yanshan University,Qinhuangdao 066004,China;Department of Instrument Science and Engineering,School of Electrical Engineering,Yanshan University,Qinhuangdao 066004,China)

机构地区：[1]燕山大学信息科学与工程学院电子与通讯工程系,河北秦皇岛066004 [2]燕山大学电气工程学院仪器科学与工程系,河北秦皇岛066004

出　　处：《光谱学与光谱分析》2024年第11期3064-3068,共5页Spectroscopy and Spectral Analysis

基　　金：国家自然科学基金项目(62173289)资助。

摘　　要：海面溢油是海洋污染的重要形式之一,溢油在风化迁移过程中,形成如水包油,油包水,水包油包水等乳化液,其中水包油类型的乳化液受水分子的影响较大,荧光特性不突出,导致轻质油乳化液分类的困难,因此如何进行高效识别,对污染治理有重要意义。选取常见的几种轻质油分别与海水、乳化剂按照不同配比混合配置水包油类型的轻质油乳化液。采用实验室搭建的便捷激光诱导荧光(LIF)系统探测轻质油乳化液的荧光光谱。构建麻雀搜索算法(SSA)优化支持向量机(SVM)的分类模型(以下简称SSA-SVM),实现乳化阶段溢油的分类识别。采用主成分分析法(PCA)对荧光光谱进行降维,选取累计贡献率为99%的前三个主成分作为输入,轻质油的种类作为输出;采用SSA迭代得到SVM的最优参数,并构建SSA-SVM分类模型;最后将测试集样本代入到模型中进行分类识别,识别准确率为100%。该研究同时构建了粒子群算法(PSO)优化支持向量机模型(简称PSO-SVM)和遗传算法(GA)优化支持向量机模型(简称GA-SVM)作为对照,实验结果表明,SSA-SVM相比PSO-SVM和GA-SVM,测试集的轻质油乳化液分类识别准确率同比提升1.77%和3.04%;并且适应度曲线在第2代就达到最高,优于PSO的第4代和GA的第36代,收敛速度更快。采用激光诱导荧光技术实现了水包油类型的轻质油乳化液的分类识别,推进了海面溢油区域分类探测机理的发展;提出的SSA-SVM模型,为轻质油乳化液的分类识别研究提供了新思路。Oil spills at sea are one of the important forms of Marine pollution.In weathering and migration,oil spills will form emulsions such as oil-in-water,water-in-oil,water-in-oil-in-water,and other emulsions.Among them,water molecules greatly affect oil-in-water emulsions,and their fluorescence characteristics are not prominent,making it difficult to classify and identify light oil emulsions.It has important significance for pollution control in the future.Several common light oils were selected to mix with seawater and emulsifiers in different proportions to prepare the light oil emulsion of the oil-in-water type.A convenient laser-induced fluorescence(LIF)system built in the laboratory was used to detect the fluorescence spectra of light oil emulsions.In this paper,the classification model of the sparrow search algorithm(SSA)optimized support vector machine(SVM)(from now on referred to as SSA-SVM)is constructed to realize the classification and identification of oil spill in the emulsion stage.Firstly,principal component analysis(PCA)was used to reduce the dimension of the fluorescence spectrum,and the first three principal components with a cumulative contribution rate of 99%were selected as inputs,and the type of light oil was taken as the output;after that,SSA is used to obtain the optimal parameters of SVM iteratively.Then,the SSA-SVM classification model was constructed.Finally,samples from the test set are substituted into the model for the classification identification,and the identification accuracy is 100%.In this study,the particle swarm optimization(PSO)support vector machine model(from now on referred to as PSO-SVM)and genetic algorithm optimization support vector machine model(from now on referred to as GA-SVM)were constructed at the same time as a comparison.From the experimental results,compared with the PSO algorithm and GA algorithm,the SSA algorithm improved the classification and recognition accuracy of the test set's lightweight oil emulsions by 1.77%and 3.04%year-on-year.The fitness curve reach

关 键 词：激光诱导荧光 麻雀搜索算法 轻质油乳化液 水包油 支持向量机 

分 类 号：O433.4[机械工程—光学工程]