土壤重金属污染高光谱遥感监测方法综述

Overview of Hyperspectral Remote Sensing Monitoring Method of Soil Heavy Metals

近年来土壤重金属污染日益严重,对生态环境安全和人类健康构成了严重威胁。高光谱遥感技术具有快速、宏观地获取地物信息的能力,为快速高效获取土壤重金属含量开展污染防治提供了科学手段。本文总结了基于土壤光谱分析和植被光谱分析的两类土壤重金属污染高光谱遥感监测的方法及原理,介绍了土壤重金属含量反演建模过程及建模方法。土壤光谱分析法预测重金属含量主要基于实验室土壤光谱展开,具有较高的预测精度,但很难实现大范围土壤重金属污染监测;植被光谱分析法多是基于野外实测光谱数据展开,预测精度相对较低,但较易与高光谱影像相结合进行区域土壤重金属污染预测,是今后研究的热点方向。多元逐步线性回归和偏最小二乘回归是土壤重金属含量反演应用最为广泛的建模方法,但反演模型往往缺乏普适性和稳定性,随着高光谱传感器和数据获取平台的发展,普适性更好稳定性更强的建模方法将不断出现。

Heavy metal pollution in soil is a serious threat to ecological environment and human health.Hyperspectral remote sensing technology has the ability to acquire ground object information quickly and macroscopically.It provides a scientific method for fast,efficiently and quantitatively acquiring the heavy metal contents in soil and pollution prevention.In this paper,the methods and principles of hyperspectral remote sensing monitoring of soil heavy metal pollution were summarized based on soil spectral analysis and vegetation spectral analysis.The process and modeling method of soil heavy metal content inversion were introduced.The prediction of heavy metal contents by soil spectral analysis is based on the laboratory soil spectrum,which has high prediction accuracy,but it is difficult to realize the large-scale monitoring of heavy metal pollution in soil.Vegetation spectral analysis is mostly based on field measuring spectral data,with relatively low prediction accuracy.However,it is easy to combine with hyperspectral images to predict regional soil heavy metal pollution.Therefore,it is a most popular research field in the future.Multiple stepwise linear regression and partial least squares regression are the most widely used modeling methods for inversion of soil heavy metal contents,but the inversion model often lacks universality and stability.With the development of hyperspectral sensors and acquisition platforms,modeling methods with better universality and stability will continue to emerge.