摘要
以陕西大柳塔神东煤矿区微生物复垦基地6a生沙棘人工林为研究对象,采用地物光谱仪分别测定接种丛枝菌根真菌(AMF)和不接种对照(CK)两个水平下沙棘叶片光谱反射率,同步监测相应叶绿素含量.筛选出与叶绿素含量相关性较好的光谱提取变量和常用植被指数作为因变量,利用BP神经网络和极限学习机(ELM)的建模方法反演叶绿素含量,并对反演模型进行精度评定.结果表明:接种丛枝菌根真菌6a后沙棘叶绿素含量仍持续增加,沙棘原始光谱的绿峰值降低;不同处理各光谱特征参数对叶绿素含量的相关性不同,选取对接菌植物和对照都显著相关的特征参数,获得光谱提取变量和常用植被指数均是7个.比较BP神经网络和ELM极限学习机的建模方法,ELM建模方法显示出更好的估测能力,其中对照组与接菌组以光谱指数和光谱提取变量为自变量时模型决定系数分别高达0.814和0.862,可以较好地估测出沙棘叶片叶绿素含量,为矿区沙棘长势提供理论依据,对微生物复垦后续生态效应具有较好的现实意义.
The 6-year-old seabuckthorn plantation in the microbial reclamation base of Daliuta Shendong Coal Mine in Shaanxi was taken as the research object.The ground spectrometer was used to measure the spectral reflectance of the leaves at two levels inoculated with arbuscular mycorrhizal fungi and non-inoculated control,and the corresponding chlorophyll content of sea buckthorn was also monitored.Screening out the spectral extraction variables and common vegetation indexes that have a good correlation with chlorophyll content,and using this as the dependent variable to invert the chlorophyll content using the modeling method of BP neural network and extreme learning machine,and assessing the accuracy of the inversion model finally.The results show that the chlorophyll content of Hippophaerhamnoidescontinue to increase 6years after inoculation with arbuscular mycorrhizal fungi,and the green peak of the original spectrum of Hippophae rhamnoides decreases.Correlation of each spectral characteristic parameter to chlorophyll content in different treatments is different.By selecting the characteristic parameters which are significantly related to the plants,the spectral extraction variables and commonly used vegetation indexes are seven.Comparing the modeling methods of BP neural network and extreme learning machine,the ELM modeling method shows better estimation ability.Among the control group and the inoculation group,with the spectral index and the spectral extraction variables as the independent variable,the model determination coefficients are up to 0.814and 0.862respectively.The ELM modeling method can estimate the chlorophyll content of seabuckthorn leaves,and provide a theoretical basis for the growth of seabuckthorn in mining areas.It has good practical significance for revealing the subsequent ecological effects of microbial reclamation.
作者
毕银丽
龚云丽
杨惠惠
BI Yinli;GONG Yunli;YANG Huihui(College of Geology and Environment,Xi’an University of Science and Technology,Xi’an,Shaanxi 710054,China;College of Geoscience and Surveying Engineering,China University of Mining and Technology(Beijing),Beijing 100083,China)
出处
《中国矿业大学学报》
EI
CAS
CSCD
北大核心
2021年第1期190-196,共7页
Journal of China University of Mining & Technology
基金
国家自然科学基金项目(51974326)
首都科技领军人才项目(Z18110006318021)。
关键词
叶绿素
AM真菌
光谱参数
BP神经网络
极限学习机
chlorophyll
AM fungi
spectral parameters
BP neural network
extreme learning machine