文献类型：期刊文献

中文题名：芦苇粉大尾蚜虫害下芦苇叶绿素高光谱反演估算

英文题名：Estimation of the hyperspectral inversion of reed chlorophyll under Hyalopterus pruni attack

作者：窦志国[1,2] 崔丽娟[1,2] 武高洁[1,2] 李晶[1,2] 潘旭[1,2] 蔡张杰[1,2] 雷茵茹[1,2] 李伟[1,2]

第一作者：窦志国

机构：[1]中国林业科学研究院湿地研究所湿地生态功能与恢复北京市重点实验室;[2]北京汉石桥湿地生态系统国家定位观测研究站

年份：2018

卷号：37

期号：10

起止页码：3163-3170

中文期刊名：生态学杂志

外文期刊名：Chinese Journal of Ecology

收录：CSTPCD;;Scopus;北大核心:【北大核心2017】；CSCD:【CSCD2017_2018】；

基金：中央级公益性科研院所基本科研业务费专项(CAFYBB2017MA028)资助

语种：中文

中文关键词：高光谱;红边;全波段;叶绿素含量;芦苇;虫害

外文关键词：hyperspectrum;red edge;full-wave band;chlorophyll content;reed;insect pest.

分类号：TP311.5

摘要：叶绿素含量是表征芦苇虫害胁迫状态的一个重要指示因子。选取遭受芦苇粉大尾蚜(Hyalopterus pruni)虫害的芦苇(Phragmites australis)作为研究对象,用便携式地物光谱仪ASD FieldSpec 4测定其叶片反射率光谱,同时用SPAD-502 Plus叶绿素测量仪测定其叶绿素相对含量(SPAD),分析健康和虫害芦苇叶片高光谱反射率与叶绿素含量间的相关关系,采用一元线性回归和偏最小二乘法回归方法,建立芦苇叶绿素含量红边位置和全波段高光谱反演估算模型。结果表明:健康芦苇叶片反射率高于虫害叶片,两种叶片叶绿素含量与高光谱的相关性存在显著差异,尤其在绿光波段和近红外波段部分区域(1400~2500nm)表现明显;全波段高光谱估算模型具有较高的准确性,健康叶片回归模型的决定系数(R2)为0.965,均方根误差(RMSE)为0.813,预测偏差比率(RPD)为3.940;虫害叶片回归模型的R2为0.966,RMSE为0.989,RPD为3.855;异地验证结果进一步表明,通过高光谱数据全波段反演能较好地估算芦苇粉大尾蚜虫害下芦苇叶绿素相对含量,这也预示着利用高光谱全波段数据估算虫害植被叶绿素相对含量是可行的。
Chlorophyll content is an important indicator for the stress state of reeds. Phragmites australis affected by Hyalopterus pruni were selected as research objects. The reflectance spectra of leaves were recorded by an ASD FieldSpec 4 portable ground object spectrometer and the chlorophyll contents（ SPAD readings） were measured using a SPAD-502 Plus Chlorophyll Meter.Univariate linear regression and partial least squares（ PLS） regression were used to analyze the relationships between the spectral reflectance and the chlorophyll content of the leaves. We established two models,a full-wave-band spectrum model and a red-edge position regression model,to estimate the chlorophyll content of the reeds. The results showed that the reflectance of healthy reed leaves was higher than that of the leaves suffered by pests. The correlation of chlorophyll content with hyperspectrum was different between the two types of leaves,especially in the spectral range of 1400-2500 nm. The full-wave-band hyperspectral estimation model had high accuracy,with the determination coefficient of healthy leaves regression model being 0.965,the mean square root error（ RMSE） being 0.813,and the predicted deviation ratio（ RPD） being 3.940.The determination coefficient of pest-bearing leaves regression model was 0. 966,RMSE being0.989,and RPD being 3.855. The cross-validation and offsite-validation results further demonstrated that full-wave-band hyperspectral data inversion could estimate the relative chlorophyll content of reeds suffered by H. pruni,indicating that it was feasible to use full-wave-band hyperspectral data to estimate the relative chlorophyll content of vegetation suffered by pests.