激光诱导击穿光谱数据特征自动提取方法研究  

Study on the Automatic Extraction Method of Spectral Data Features in Laser Induced Breakdown Spectroscopy

作　　者：刘立拓[1] 刘建国[1] 赵南京[1] 鲁翠萍[1] 陈东[1] 石焕[1] 王春龙[1] 张玉钧[1] 刘文清[1] 

机构地区：[1]中国科学院安徽光学精密机械研究所,中国科学院环境光学与技术重点实验室,安徽合肥230031

出　　处：《光谱学与光谱分析》2011年第12期3285-3288,共4页Spectroscopy and Spectral Analysis

基　　金：国家自然科学基金项目(60908018);中国科学院知识创新工程重大项目(kzcx-1yw-06-01);安徽省科技计划项目(083H3b1581);所长基金项目(073H371141)资助

摘　　要：根据激光诱导击穿光谱(LIBS)展宽主要为非线性Lorentz函数模型,利用Levenberg-Marquardt(L-M)算法对该理论模型中的待定参数进行优化估值,扣除了光谱信号中的连续背景,校正和还原了峰位和峰强,并对拟合数据点自动提取方法进行了分析;以MatrixVB与VB混合编程相结合,实现了基于L-M算法的LIBS光谱数据自动拟合与特征参数提取,与手动操作的常用数据分析软件Origin7.5的拟合结果进行了比较,验证了该方法的稳定性与可靠性。According to that the spectral line broadening of laser induced breakdown spectroscopy（LIBS）was lorentz nonlinearfunction model,using Levenberg-Marquardt（L-M）agorithm to optimally estimate the undetermined parameter in the function,the continuum background radiation and peak position and peak intensity were corrected,and the method of data points automaticextraction for fitting was analysed.Automatic fitting and feature parameter extraction for spectral data of LIBS based on L-M al-gorithm that was hybrid programming with matrixVB and VB were accomplished,and the result was compared with the origin7.5data analysis software which was manual and commonly used and verified its stability and reliability.

关 键 词：激光诱导击穿光谱 特征提取 LEVENBERG-MARQUARDT算法 非线性拟合 

分 类 号：O433.5[机械工程—光学工程]