光谱学与光谱分析, 2017, 37 (5): 1649, 网络出版: 2017-06-20
山东省焦家式金矿区土壤重金属铬高光谱监测研究
Concentration Estimation of Heavy Metal Cr in Soilsfrom Jiaojia-Type Gold Mining Areas Using Reflectance Spectroscopy
矿区 光谱技术 铬 预测模型 空间分布 Goldmining area Reflectance spectroscopy Cr Predictive model Spatial distribution
摘要
矿区土壤环境问题日益突出, 以山东省烟台市焦家式金矿区为典型区, 系统采集了85个表层土壤样品进行野外光谱实测, 其中35份用于重金属铬含量测定。 利用光谱变换技术对原始光谱提取了一阶微分、 二阶微分和去包络线三种光谱指标, 采用偏最小二乘回归法建立光谱数据与铬含量间的定量关系, 将最优预测模型应用于其余50个样本点铬浓度值的获取, 最终通过克里金插值实现了对研究区土壤重金属铬的快速监测。 结果表明, 基于一阶微分数据模型预测精度最高, 其次为二阶微分, 原始光谱和去包络线数据。 研究区铬的含量与金矿的分布情况密切相关, 即金矿密集区域铬含量更高, 这表明金矿开采对土壤中铬的含量与分布存在一定影响。
Abstract
Heavy metal contamination has been rising markedly in mining area and a rapid and accurate assessment technique is of utmost importance. In this study, visible and near-infrared spectra combined with partial least square regression (PLSR) are developed to achieve the rapid monitoring of Cr in soilsfrom Jiaojia-type gold mines. A total of eighty five soil samples were collected to measure the spectra on the spot, thirty five of which were used for chemical analysis. PLSR models were established using primitive spectra (R) and first derivative (RD1), second derivative (RD2) and continuum removal (CR) data to find the most suitable calibration. The best predictive model that used RD1 spectra was applied to determine the contents of Cr in the untested samples and the spatial distribution map was created by universal kriging interpolation. Results illustrated that the major abundance of Cr was concentrated in the range of 36~48 mg·kg-1, with the highest value being 71.8 mg·kg-1 and lowest being 20.9 mg·kg-1. Furthermore, elevation and the orefield showed prominent impacts on the content values of Cr. Generally plain and the areas closed to gold mines tend to have higher concentrations of Cr.
王菲, 王集宁, 曹文涛, 康日斐, 曹见飞, 吴泉源. 山东省焦家式金矿区土壤重金属铬高光谱监测研究[J]. 光谱学与光谱分析, 2017, 37(5): 1649. WANG Fei, WANG Ji-ning, CAO Wen-tao, KANG Ri-fei, CAO Jian-fei, WU Quan-yuan. Concentration Estimation of Heavy Metal Cr in Soilsfrom Jiaojia-Type Gold Mining Areas Using Reflectance Spectroscopy[J]. Spectroscopy and Spectral Analysis, 2017, 37(5): 1649.