基于核偏最小二乘的矿区土壤Cu含量高光谱反演
|
摘要
本文探究应用高光谱遥感手段反演铜锌矿区土壤Cu含量的可行性,以湖南省某矿区土壤为例。在对原始高光谱重采样、一阶微分、对数、连续统预处理后,分别进行与Cu含量的相关性分析,最终选用一阶微分变换光谱数据进行建模。在建模反演时,针对多元线性回归(MFL)和传统偏最小二乘(PLS)在应用过程中没有考虑变量间的非线性关系的缺点,提出了基于核偏最小二乘(KPLS)回归的土壤Cu含量预测模型。研究结果表明,相对于PLS和MFL,KPLS能较好的提升土壤Cu含量估算能力,预测样本的平均相对误差为13.25%,明显高于MFL的32.22%和PLS的14.18%。研究结果也表明了高光谱遥感手段可以反演矿区土壤Cu含量,且核偏最小二乘模型也可为其它土壤重金属的反演提供参考。
Taking a mining area soil in Hunan Prov ince as an example, we explored the feasibility of using hyperspectral remote sensing to invert Cu content in copper-zinc mines. The original spectrum was pretreated including re-sampling, first-order differential, logarithmic, and continuum, and then their correlation analysis with Cu content was performed, finally, first-order differential transformation spectral data were selected to construct the models. In modeling inversion, we did not consider the disadvantages of non-linear relationships between variables by using the methods of multi-factor linear regression(MFL) and traditional partial least squares(PLS), and proposed the prediction model for soil Cu content based on kernels partial least squares(KPLS) regression. The results showed that compared with traditional MFL and PLS, KPLS better enhanced the estimation ability of soil Cu content, and the average relative error of the prediction sample was 13.25%, significantly 32.22% higher than that of MFL and 14.18%higher than that of PLS. The hyperspectral remote sensing method can be used to invert Cu content in the mining area,and the kernels partial least squares model can also provide a reference for inversion of the other heavy metals.
Taking a mining area soil in Hunan Prov ince as an example, we explored the feasibility of using hyperspectral remote sensing to invert Cu content in copper-zinc mines. The original spectrum was pretreated including re-sampling, first-order differential, logarithmic, and continuum, and then their correlation analysis with Cu content was performed, finally, first-order differential transformation spectral data were selected to construct the models. In modeling inversion, we did not consider the disadvantages of non-linear relationships between variables by using the methods of multi-factor linear regression(MFL) and traditional partial least squares(PLS), and proposed the prediction model for soil Cu content based on kernels partial least squares(KPLS) regression. The results showed that compared with traditional MFL and PLS, KPLS better enhanced the estimation ability of soil Cu content, and the average relative error of the prediction sample was 13.25%, significantly 32.22% higher than that of MFL and 14.18%higher than that of PLS. The hyperspectral remote sensing method can be used to invert Cu content in the mining area,and the kernels partial least squares model can also provide a reference for inversion of the other heavy metals.
引文
[1]张东辉,赵英俊,陆冬华,等.高光谱在土壤重金属信息提取中的应用与实现[J].土壤通报,2018,49(01):31-37.
[2]郭云开,周烽松,丁美青,等.水稻冠层与土壤高光谱反演土壤重金属对比研究[J].遥感信息,2017,32(02):173-179.
[3]贾亚琪,程志飞,刘品祯,等.煤矿区周边农田土壤重金属积累特征及生态风险评价[J].土壤通报,2016,47(02):474-479.
[4]易凌霄,曾清如.洞庭湖区土壤重金属污染现状及防治对策[J].土壤通报,2015,46(06):1509-1513.
[5]BOOCA B,ALIMONTIN A,PETRUCCI F,et al.Quantification of trace elements by sector field inductively coupled plasma mass spectrometry in urine,serum,blood and cerebrospinal fluid of patients with Parkinson's disease[J].Spectrochimica Acta Part B,2004,59(4):559-557.
[6]KOOISTRA L,WEHRENS R,LEUVEN R,et al.Possibilities of visible-near-infrared spectroscopy for the assessment of soil contamination in river floodplains[J].Analytica Chimica Acta,2001,446(1-2):97-105.
[7]KEMPER T,SOMMER S.Estimate of heavy metal contamination in soils after a mining accident using reflectance spectroscopy[J].Environmental Science&Technology,2002,36(12):2742-2747.
[8]王维,沈润平,吉曹翔.基于高光谱的土壤重金属铜的反演研究[J].遥感技术与应用,2011,26(03):348-354.
[9]吕杰,徐静,闫振国.基于小波神经网络的矿区土壤铜含量反演研究[J].矿业研究与开发,2015,35(04):68-70.
[10]涂宇龙,邹滨,姜晓璐,等.矿区土壤Cu含量高光谱反演建模[J].光谱学与光谱分析,2018,38(02):575-581.
[11]李媛媛,李微,刘远,等.基于高光谱遥感土壤有机质含量预测研究[J].土壤通报,2014,45(06):1313-1318.
[12]郭云开,刘宁,刘磊,等.土壤Cu含量高光谱反演的BP神经网络模型[J].测绘科学,2018,43(01):135-139.
[13]黄长平,刘波,张霞,等.土壤重金属Cu含量遥感反演的波段选择与最佳光谱分辨率研究[J].遥感技术与应用,2010,25(03):353-357.
[14]ROSIPAL R.Kernel partial least squares for nonlinear regression and discrimination[J].Neural Network World,2003.13(3):291-300.
[15]ROSIPAL R,TREJO L J.Kernel partial least squares regression in reproducing kernel hilbert space[J].Journal of machine learning research,2001,2(Dec):97-123.
[16]BENNETT K P,EMBRECHTS M J.An optimization perspective on kernel partial least squares regression[J].Nato Science Series sub series III computer and systems sciences,2003,190:227-250.
[17]张曦,陈世和,陈锐民,等.基于核偏最小二乘的电厂热力参数预测与估计[J].中国电机工程学报,2011,31(S1):193-199.
[18]姚林,阳建宏,何飞,等.基于核偏最小二乘的锌层重量预测模型[J].控制工程,2008(02):154-157.
[2]郭云开,周烽松,丁美青,等.水稻冠层与土壤高光谱反演土壤重金属对比研究[J].遥感信息,2017,32(02):173-179.
[3]贾亚琪,程志飞,刘品祯,等.煤矿区周边农田土壤重金属积累特征及生态风险评价[J].土壤通报,2016,47(02):474-479.
[4]易凌霄,曾清如.洞庭湖区土壤重金属污染现状及防治对策[J].土壤通报,2015,46(06):1509-1513.
[5]BOOCA B,ALIMONTIN A,PETRUCCI F,et al.Quantification of trace elements by sector field inductively coupled plasma mass spectrometry in urine,serum,blood and cerebrospinal fluid of patients with Parkinson's disease[J].Spectrochimica Acta Part B,2004,59(4):559-557.
[6]KOOISTRA L,WEHRENS R,LEUVEN R,et al.Possibilities of visible-near-infrared spectroscopy for the assessment of soil contamination in river floodplains[J].Analytica Chimica Acta,2001,446(1-2):97-105.
[7]KEMPER T,SOMMER S.Estimate of heavy metal contamination in soils after a mining accident using reflectance spectroscopy[J].Environmental Science&Technology,2002,36(12):2742-2747.
[8]王维,沈润平,吉曹翔.基于高光谱的土壤重金属铜的反演研究[J].遥感技术与应用,2011,26(03):348-354.
[9]吕杰,徐静,闫振国.基于小波神经网络的矿区土壤铜含量反演研究[J].矿业研究与开发,2015,35(04):68-70.
[10]涂宇龙,邹滨,姜晓璐,等.矿区土壤Cu含量高光谱反演建模[J].光谱学与光谱分析,2018,38(02):575-581.
[11]李媛媛,李微,刘远,等.基于高光谱遥感土壤有机质含量预测研究[J].土壤通报,2014,45(06):1313-1318.
[12]郭云开,刘宁,刘磊,等.土壤Cu含量高光谱反演的BP神经网络模型[J].测绘科学,2018,43(01):135-139.
[13]黄长平,刘波,张霞,等.土壤重金属Cu含量遥感反演的波段选择与最佳光谱分辨率研究[J].遥感技术与应用,2010,25(03):353-357.
[14]ROSIPAL R.Kernel partial least squares for nonlinear regression and discrimination[J].Neural Network World,2003.13(3):291-300.
[15]ROSIPAL R,TREJO L J.Kernel partial least squares regression in reproducing kernel hilbert space[J].Journal of machine learning research,2001,2(Dec):97-123.
[16]BENNETT K P,EMBRECHTS M J.An optimization perspective on kernel partial least squares regression[J].Nato Science Series sub series III computer and systems sciences,2003,190:227-250.
[17]张曦,陈世和,陈锐民,等.基于核偏最小二乘的电厂热力参数预测与估计[J].中国电机工程学报,2011,31(S1):193-199.
[18]姚林,阳建宏,何飞,等.基于核偏最小二乘的锌层重量预测模型[J].控制工程,2008(02):154-157.