摘要
胡杨叶绿素含量的快速、无损、准确监测,可以为荒漠河岸林的健康遥感监测提供理论依据。本研究以胡杨作为研究对象,测得不同地下水埋深区域的胡杨叶片进行光谱与之相对应的叶绿素含量,分析不同埋深区叶片的叶绿素含量及光谱特征变化,采用灰色关联度与相关系数法分析所光谱与叶绿素的关系,提取出敏感波段后,利用机器学习算法,建立基于BP神经网络、PLSR、SVM的叶绿素含量的光谱估算模型。结果显示:随地下水埋深的增大,叶绿素含量呈先上升后下降趋势,不同埋深下胡杨叶绿素含量高低依次是:4~6 m>2~4 m>6~8 m>8~10 m>0~2 m,说明叶绿素含量对地下水埋深变化较敏感;光谱反射率曲线在峰谷处的反射率大小依次是4~6 m>0~2 m>6~8 m>8~10 m>2~4 m,说明光谱曲线对地下水埋深梯度变化有响应。一阶导数的BP神经网络估算模型具有较好的稳定性及预测能力,其决定系数(R^(2))、均方根误差(RMSE)、相对分析误差(RPD)分别为0.81,0.20和2.14。研究表明:基于一阶导数的BP神经网络估算模型对叶绿素a+b具有较好的预测能力,可以用于快速准确地监测胡杨叶片的叶绿素含量,为荒漠绿洲过渡带的生态保育提供技术支撑。
As an important indicator of the ecological environment in arid areas,Populus has the role of inhibiting desertification processes,curbing wind and sand and maintaining the balance of fragile ecosystems.Chlorophyll content is a key indicator of plant health,and rapid and accurate monitoring of chlorophyll content in poplar leaves plays an important role in ecological conservation in the desert oasis transition zone.Therefore,this study takes gray poplar as the research object,performs spectral testing and sample collection of poplar leaves in different groundwater burial depth areas,and uses machine learning algorithms to establish a spectral estimation model based on chlorophyll content of BP neural network,PLSR and SVM,and the results show that the BP neural network estimation model of first derivative has good stability and prediction ability,and its decision coefficient(R^(2)),root mean square error(RMSE),and relative analysis error(RPD)are established.0.81,0.20 and 2.14,respectively,showed good predictive power for total chlorophyll,while the RPDs of other models were less than 2.0.These results provide certain technical support for the healthy remote sensing monitoring of desert riverbank forests,and also provide a theoretical reference for the construction of ecological environment and vegetation restoration in arid areas.
作者
温云梦
张冬冬
王家强
多晶
蔡海辉
柳维扬
WEN Yun-meng;ZHANG Dong-dong;WANG Ja-qiang;DUO Jing;CAI Hai-hui;LIU Wei-yang(Agricultural College,Tarim University,Alar Xinjiang 843300,P.R.China)
出处
《西部林业科学》
CAS
北大核心
2022年第4期87-95,共9页
Journal of West China Forestry Science
基金
国家自然科学基金项目“荒漠河岸林胡杨叶片对地下水埋深的光谱响应研究”(31860172)。
关键词
高光谱
叶绿素
估算模型
胡杨
干旱区
hyperspectral
chlorophyll
estimation model
Populus euphratica
arid area