论文部分内容阅读
尾矿重金属污染是当今矿区环境污染面临的主要问题之一,精确反演土壤重金属含量对矿区土壤污染监测和治理具有非常重要的意义。以陕西金堆城矿区尾矿为研究区,利用ASD光谱仪测量土壤光谱,通过实验室化学分析获取土壤样本铜元素含量;将Isomap流形学习方法应用于土壤高光谱数据降维,利用随机森林方法对矿区尾矿土壤的Cu含量进行反演建模,并与原始高光谱数据反演结果和PCA降维后的反演结果进行对比。结果表明:土壤铜含量反演模型在经过Isomap降维后的光谱数据集上预测铜元素含量的相关系数R2为0.7272,均方根误差RMSE为1140.20,在预测的准确性方面均优于原始高光谱数据。研究结果为探索土壤高光谱数据特征提取提供了理论依据,同时对尾矿重金属污染监测具有重要的现实指导意义。
Tailings heavy metal pollution is one of the main problems in environmental pollution in mining areas today. Accurately inverting heavy metal content in soils is of great significance to the monitoring and management of soil pollution in mining areas. Taking the tailings of Jinduicheng mine in Shaanxi Province as the research area, the spectrum of soil was measured by ASD spectrometer, and the content of copper in soil samples was obtained by laboratory chemical analysis. The learning method of Isomap manifold was applied to reduce the dimension of hyperspectral data of soil. The Cu content of the tailing soil in the mining area was inversely modeled and compared with the original hyperspectral data inversion results and PCA dimensionless inversion results. The results showed that the correlation coefficient R2 of copper content in the soil copper content inversion model after the Isomap dimensionality reduction was 0.7272 and the root mean square error RMSE was 1140.20, both of which were better than the original ones in predicting the accuracy Spectral data. The results provide a theoretical basis for exploring the feature extraction of soil hyperspectral data, meanwhile, it has important practical significance for the monitoring of tailings heavy metal pollution.