摘要
不同种类土壤中常量元素含量不同,对应的激光诱导击穿光谱(LIBS)特征谱线强度不同,以此为分类依据,采用LIBS技术和支持向量机(SVM)相结合研究土壤分类问题。对八种不同种类土壤进行研究,并基于SVM建立土壤分类模型。采集土壤中六种常量元素Fe、Si、Ti、Ca、Mg、K的LIBS光谱,将上述元素谱线的峰值强度归一化后作为特征变量,划分训练集与测试集进行分类模型训练。八种土壤分类准确率的算术平均值达到95.625%,表明LIBS技术用于土壤分类具有较好的效果。LIBS技术具有快速检测、多元素同时分析等优点,为土壤分类和土壤普查提供一种新的可行方法。
Different types of soils have different contents of macronutrients,corresponding to different intensity of laser-induced breakdown spectra(LIBS)characteristic spectral lines,which are used as a basis for classification,and a combination of LIBS technique and sup-port vector machine(SVM)is used to study the soil classification.Eight different types of soils are studied and a soil classification model is developed based on SVM.The LIBS spec-tra of six constant elements Fe,Si,Ti,Ca,Mg and K are collected from the soils,and the peak intensities of the above elements are normalized as the feature variables,and the train-ing and test sets are divided for training the classification models.The arithmetic mean of the classification accuracy of the eight soils is 95.625%,indicating that the LIBS technique has good results for soil classification.The LIBS technique has the advantages of rapid detection and simultaneous analysis of multiple elements,which provides a new feasible method for soil classification and soil census.
作者
刘志忠
赵吉
周晨阳
赵静怡
李倩
李业秋
LIU Zhizhong;ZHAO Ji;ZHOU Chenyang;ZHAO Jingyi;LI Qian;LI Yeqiu(Shenyang Ligong University,Shenyang 110159,China;Jilin Province key Laboratory of Measuring Instruments and Technology,Changchun 130103,China)
出处
《沈阳理工大学学报》
CAS
2023年第6期89-94,共6页
Journal of Shenyang Ligong University
基金
辽宁省博士科研启动基金计划项目(2021-BS-161)
辽宁省教育厅科学研究面上重点项目(LJKZ0262)。
关键词
激光诱导击穿光谱
支持向量机
土壤分类
laser-induced breakdown spectroscopy
support vector machine
soil classification
