激光诱导击穿光谱结合间隔偏最小二乘法检测土壤中的Pb  

Determination of Pb in Soil by Laser Induced Breakdown Spectroscopy Coupled with Interval Partial Least Squares

作　　者：胡慧琴[1] 黄林[1] 涂建平[1] 徐雪红[1] 姚明印[1] 陈添兵[1] 刘木华[1] 杨平[1] 王彩虹[1] 

机构地区：[1]江西农业大学生物光电及应用实验室,江西南昌330045

出　　处：《应用激光》2015年第1期104-109,共6页Applied Laser

基　　金：国家自然科学基金资助项目(项目编号:31460419);国家自然科学基金资助项目(项目编号:31271612);江西省重大科技资助项目(项目编号:20143ACB21013);江西省教育厅科技计划资助项目(项目编号:GJJ10681);江西省学术带头人计划资助项目(项目编号:09004004);赣鄱"555"英才;大学生创新计划资助项目(项目编号:201410410010)

摘　　要：为了提高激光诱导击穿光谱（LIBS）技术检测自然土壤中Pb的检测精度,提出采用间隔偏最小二乘法（IPLS）定量预测模型。对土壤在400~417nm波段的特征光谱进行平滑预处理后,建立偏最小二乘法（PLS）定量模型,得到训练集的相关系数为0.974 2,且斜率为0.983。建立IPLS模型时,把所选波段均分成了25个子区间,得到第八个子区间包含了Pb的特征光谱405.78nm,且交叉验证均方根误差最小,选择该区间建立模型得到训练集相关系数为0.985 3,斜率为1.121。预测集中,土壤样品Pb的真实浓度与预测浓度之间的相对误差在13%以内,平均相对误差为7.00%。研究表明IPLS法应用于LIBS定量检测土壤中的Pb是可行的,且该定量模型预测效果优于PLS法。In order to improve the detection accuracy in detecting Pb of soil in the nature better by laser induced breakdown spectroscopy（LIBS）,A quantitative analysis model based on interval partial least squares（IPLS）in LIBS technique is adopted.After the spectrum data of soil in 400~417nm is preprocessed by smoothing,quantitative model is established.The correlation coefficient of training set in PLS model is 0.974 2.and the slope is 0.983.However,when the IPLS model is established,the selected spectrum is divided into 25sub-intervals,the eighth subinterval contains the characteristic line of Pb 405.78 nm,and the Root Mean Square Error Cross-Validation（RMSECV）is the minimum.the correlation coefficient of training IPLS model established by the eighth subinterval is 0.985 3,and the slope is 1.121.The relative error between predicted and actual concentration of Pb is below 13%in the prediction set,and the average relative error is 7.00%.It is demonstrated that it is feasible for detecting Pb of soil by LIBS,and IPLS method is prior than PLS method.

关 键 词：LIBS 定量分析 IPLS 重金属 土壤 

分 类 号：O657.3[理学—分析化学]