川西铜矿周边土壤铬含量高光谱反演研究

Study on hyperspectral inversion of chromium content in soil around the West Sichuan copper mine

王光羽 ¹杨斌 ²陈卓尔 ¹魏添翼 ¹杨坤 ¹卓思杰¹

扫码查看

作者信息

1. 西南科技大学环境与资源学院,四川绵阳 621010
2. 西南科技大学环境与资源学院,四川绵阳 621010;国家遥感中心绵阳科技城分部,四川绵阳 621010;西南科技大学四川天府新区创新研究院,成都 621010
折叠

摘要

矿产资源的开采,给其周边土壤带来许多环境问题,如何快速筛查出矿区周边土壤污染情况尤为重要.本文以川西某铜矿周边土壤作为研究对象,采用 LASSO 算法对 SG 平滑后的高光谱数据 RSG和经过多元散射校正(MSC)、一阶微分(FD)、倒数变换(RT)的光谱数据进行特征波段的筛选.使用偏最小二乘回归(PLSR)、随机森林(RF)、支持向量机回归(SVR)、极端梯度提升(XGBoost)、反向传播神经网络(BPNN)5 种模型进行反演.结果表明,RSG、MSC、FD筛选出的特征波段集中在近红外区域,RT 筛选出的特征波段集中在可见光区域;MSC-SVR 模型的R2,RMSE和 RPD分别为 0.763、6.745、2.06,在所有模型中精度最高,该模型可用于研究区土壤中铬的快速监测.

Abstract

The extraction of mineral resources has brought many environmental problems to the surrounding soil,and it is particularly important to quickly screen for soil pollution in the mining area.This article takes the soil around a copper mine in western Sichuan as the research object.Using LASSO algorithm to screen feature bands for SG smoothed hyperspectral data RSG and spectral data after multivariate scattering correction(MSC),first-order differentiation(FD),and reciprocal transformation(RT).Perform inversion using five models:Partial Least Squares Regression(PLSR),Random Forest(RF),Support Vector Machine Regression(SVR),eXtreme Gradient Boosting(XGBoost),and Backpropagation Neural Network(BPNN).The results show that the feature bands selected by RSG,MSC,and FD are concentrated in the near-infrared region,while the feature bands selected by RT are concentrated in the visible light region;the R2,RMSE,and RPD of the MSC-SVR model are 0.763,6.745,and 2.06,respectively,with the highest accuracy among all models.This model can be used for rapid monitoring of chromium in the study area.

关键词

高光谱反演/土壤重金属含量/光谱分析/川西铜矿/铬含量

Key words

hyperspectral inversion/soil heavy metal content/spectral analysis/West Sichuan copper mine/chromium content

引用本文复制引用

出版年

2025

测绘工程

黑龙江工程学院中国测绘学会

测绘工程

影响因子：1.78

ISSN：1006-7949

段落导航