光学学报, 2013, 33 (3): 0330002, 网络出版: 2013-01-16
基于支持向量机回归的水体重金属激光诱导击穿光谱定量分析研究
Quantitative Analysis of Laser-Induced Breakdown Spectroscopy of Heavy Metals in Water Based on Support-Vector-Machine Regression
光谱学 激光诱导击穿光谱 支持向量机回归 重金属 石墨 spectroscopy laser-induced breakdown spectroscopy support-vector-machine regression heavy metals graphit
摘要
建立了基于自适应核的激光诱导击穿光谱支持向量机回归定量分析模型。通过石墨富集、洛伦兹拟合和碳内标归一化,增强等离子体信号强度,减小环境噪声和能量抖动对水体重金属浓度测量的影响。实现了基于支持向量机回归智能算法的激光诱导击穿光谱定量分析。铅铜的平均相对标准偏差分别为6.4361%和6.9291%,最大相对标准偏差分别为9.1009%和8.9280%,平均相对误差分别为1.6765%和1.2478 %,最大相对误差分别为5.5759%和4.2604%,相关系数分别为0.9979和0.9997。该研究为进一步实现水中痕量金属元素的快速定量分析提供了方法和数据参考。
Abstract
The quantitative analysis model of laser-induced breakdown spectroscopy with adaptive kernel is established. Effect of ambient noise and energy jitter in measured density of heavy metals is gradually removed by Lorentz fitting and carbon normalization, and the intensity of plasmas is enhanced by graphite enrichment. Quantitative analysis of laser-induced breakdown spectroscopy based on regression intelligent algorithm of support vector machine is achieved. The average relative standard deviations of lead and copper are 6.4361% and 6.9291%, and the maximum standard deviations are 9.1009% and 8.9280%.The average relative errors of lead and copper are 1.6765% and 1.2478 %, and the maximum relative errors are 5.5759% and 4.2604%. The correlation coefficients of lead and copper are 0.9979 and 0.9997. Methods and reference data are provided for the further study of fast measurement of trace heavy metals in water by laser-induced breakdown spectroscopy.
王春龙, 刘建国, 赵南京, 马明俊, 王寅, 胡丽, 张大海, 余洋, 孟德硕, 章炜, 刘晶, 张玉钧, 刘文清. 基于支持向量机回归的水体重金属激光诱导击穿光谱定量分析研究[J]. 光学学报, 2013, 33(3): 0330002. Wang Chunlong, Liu Jianguo, Zhao Nanjing, Ma Mingjun, Wang Yin, Hu Li, Zhang Dahai, Yu Yang, Meng Deshuo, Zhang Wei, Liu Jing, Zhang Yujun, Liu Wenqing. Quantitative Analysis of Laser-Induced Breakdown Spectroscopy of Heavy Metals in Water Based on Support-Vector-Machine Regression[J]. Acta Optica Sinica, 2013, 33(3): 0330002.