支持向量机结合FTIR的沥青混合料老化程度鉴别  

Identification of asphalt mixture aging degree by SVM combined with FTIR

作　　者：朱怡烁 张维 胡锦江 ZHU Yishuo;ZHANG Wei;HU Jinjiang(School of Machinery and Automation,Wuhan University of Science and Technology,Wuhan 430081,China;Hubei Institute of Measurement and Testing Technology,Wuhan 430223,China)

机构地区：[1]武汉科技大学机械自动化学院,湖北武汉430081 [2]湖北省计量测试技术研究院,湖北武汉430223

出　　处：《传感器与微系统》2025年第4期74-77,82,共5页Transducer and Microsystem Technologies

基　　金：国家自然科学基金资助项目(51778509,61901310)。

摘　　要：为实现沥青混合料老化程度的分类识别,本文基于傅里叶变换红外(FTIR)光谱技术,采用无信息变量消除(UVE)方法结合浣熊优化算法(COA)优化支持向量机(SVM),建立了分类识别模型。首先,采集3种不同老化程度的沥青混合料红外光谱数据,并运用S-G平滑+标准正态变量(SNV)变换对原始光谱进行预处理;再用UVE算法减少光谱冗余信息,从7157个变量中获得了1197个变量;最后引入COA对SVM惩罚因子C和核函数半径σ优化,建立识别模型,并与粒子群优化(PSO)算法、鲸鱼优化算法(WOA)对SVM优化效果进行对比。结果表明:经UVE进行光谱变量筛选明显提高了模型精度,UVE-COA-SVM训练集和测试集正确率均为100%,优于UVE-PSO-SVM和UVE-WOA-SVM,该方法可用于沥青混合料老化程度识别模型的建立。In order to achieve classification and recognition of the aging degree of asphalt mixtures,based on Fourier transform infrared(FTIR)spectroscopy technology,the uninformative variable elimination(UVE)method is combined with the Raccoon optimization algorithm(COA)to optimize the support vector machine(SVM)and establish a classification identification model.Firstly,IR spectroscopy data of three different aging degrees of asphalt mixtures are collected,and the original spectrum are preprocessed using S-G smoothing+standard normal variate(SNV)transform.Then,the UVE algorithm is employed to reduce spectral redundant information,1197 variables from the original 7157 variables is obtained.Finally,the COA is introduced to optimize the SVM penalty factor C and the kernel function radiusσand establish the recognition model.The optimization effects of SVM using particle swarm optimization(PSO)and whale optimization algorithm(WOA)are compared.The results indicate that the spectral variable screening through UVE significantly improves the model precision.The UVE-COA-SVM achieves 100%accuracy on both the training and testing sets,which is better than UVE-PSO-SVM and UVE-WOA-SVM.This method can be used to establish a model for the identification of aging degree of asphalt mixtures.

关 键 词：沥青混合料 傅里叶变换红外光谱 浣熊优化算法 支持向量机 老化识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] O657.33[自动化与计算机技术—计算机科学与技术]