摘要
利用高光谱遥感技术监测并识别农作物受重金属污染信息是当今热点,研究设置了不同浓度铜离子(Cu^(2+))、铅离子(Pb^(2+))胁迫梯度的玉米盆栽实验,并测取了玉米叶片的光谱及叶片中重金属离子与叶绿素含量。基于获取的光谱数据,将光谱划分为紫谷、蓝边、绿峰、红谷、红边和红肩六个光谱特征区间,通过光谱的一阶微分和二维多重信号分类(2D-MUSIC)算法构造空间谱,对各光谱特征区间进行变换分析。实验结果表明:蓝边、绿峰和红边阵列信号的空间谱在Cu^(2+)胁迫下为双高峰,在Pb^(2+)胁迫下为单高峰,以此能够快速、直观地区分玉米叶片所受重金属污染的Cu^(2+)和Pb^(2+)元素类别。红谷和红肩阵列信号空间谱的方位角谱峰值与玉米叶片中Cu^(2+)含量的相关系数分别达到-0.954 5和-0.964 8,说明用于监测Cu^(2+)污染程度时效果理想;紫谷阵列信号空间谱的方位角谱峰值与玉米叶片中Pb^(2+)含量的相关系数达到-0.999 8,说明用于监测Pb^(2+)污染程度时效果理想。同时通过与常规重金属污染监测方法绿峰高度(GH)、红边位置(REP)、红边最大值(MR)、红边一阶微分包围面积(FAR)的应用结果进行比较分析,空间谱法的应用结果与玉米叶片中重金属离子含量的相关性较高,从而验证了空间谱应用于玉米重金属污染信息监测具有更好的有效性和优越性。
Identifying and monitoring the heavy metal pollution information of crops isthe research focus by hyperspectral remote sensing technology today.The potted corn experiments were set up with different Cu^2＋and Pb^2＋stress gradients in this research,measuring the spectral data,the content of heavy metals icon and chlorophyll of corn leaves.On the basis of the collected data,the spectra were divided into six spectral characteristic intervals：purple vallley,blue edge,green peak,red valley,red edge and red shoulder,and spectral characteristic intervals were transformed and analyzed by spatial spectrum,which was constructed by first order differential and 2 D multiple signal classification（2D-MUSIC）algorithm.The analyzed and processed results show,the spatial spectra of the array signals of the blue edge,green peak and red edge are double peaks under Cu^2＋stress.However,the spatial spectra of the array signals of the blue edge,green peak and red edge were single peak under Pb^2＋stress.Thus,the heavy metals elements categories of Cu^2＋and Pb^2＋in polluted corn could be quickly and visually distinguished.Azimuth spectrum peaks of array signal spatial spectra of red valley and red shoulder were gradually decreased under Cu^2＋stress,and the correlation coefficients of azimuth spectrum peak values of red valley and red shoulder and the Cu^2＋contents in corn leaves reached-0.954 5 and-0.964 8.It was indicated that the effect was ideal when monitoring the level of Cu^2＋pollution;azimuth spectrum peaks of array signal spatial spectrum of purple valley were gradually decreased under Pb^2＋stress,and the correlation coefficient of azimuth spectrum peak value of purple vallley and the Pb^2＋contents in corn leaves reached 0.999 8,it was indicated that the effect was ideal when the level of Pb^2＋pollution was monitored.At the same time,the application results of the spatial spectrum were analyzed and compared with the results obtained by some conventional methods such as green-peak height（GH）,red edg
作者
杨可明
张伟
王晓峰
孙彤彤
程龙
YANG Ke-ming;ZHANG Wei;WANG Xiao-feng;SUN Tong-tong;CHENG Long(College of Geoscience and Surveying Engineering,China University of Mining ＆ Technology（Beijing）,Beijing 100083,China)
出处
《光谱学与光谱分析》
SCIE
EI
CAS
CSCD
北大核心
2018年第7期2200-2208,共9页
Spectroscopy and Spectral Analysis
基金
国家自然科学基金项目(41271436)
中央高校基本科研业务费专项资金项目(2009QD02)资助
关键词
重金属污染
玉米叶片光谱
二维多重信号分类
空间谱
光谱特征区间
Heavy metal pollution
Corn leaf spectrum
2D multiple signal classification
Spatial spectrum
Spectral characteristic intervals
