摘要
【目的】阐明稻纵卷叶螟(Cnaphalocrocis medinalis)危害孕穗期水稻后,水稻冠层、未受害展开叶和受害已卷叶片的光谱特征,建立基于光谱参数的稻纵卷叶螟危害程度的诊断模型。【方法】利用便携式光谱仪测定不同卷叶率小区水稻的冠层光谱反射率,同时在不同卷叶率小区内采集未受害展开叶和已卷叶带回实验室进行室内单叶的光谱反射率测定,并采用相关分析与回归建模方法组建稻纵卷叶螟危害程度的光谱诊断模型。【结果】水稻冠层光谱反射率在近红外光区域内随卷叶率级别的升高而降低,738—1 000 nm处的反射率可较好地表征出水稻受稻纵卷叶螟危害的程度。不同卷叶率小区内的未受害叶的光谱反射率也可很好地表征水稻的受害级别,在512—606和699—1 000 nm处的反射率与小区卷叶率级别呈极显著的负相关。已受害卷叶的反射率在582—688 nm处与受害级别呈极显著正相关。水稻受稻纵卷叶螟危害后,在冠层、未受害叶及已受害卷叶光谱的红边幅值与红边面积有明显变化。利用水稻冠层光谱的红边幅值、未受害健康叶片550 nm处的反射率建立的稻纵卷叶螟危害程度的一元回归模型的诊断误差较小,而同时利用冠层、受害叶和未受害叶光谱组建的逐步回归模型的诊断误差最小,可用于小区稻纵卷叶螟危害的监测。【结论】受害区域内水稻冠层在738—1 000 nm处和未受害叶片在512—606和699—1 000 nm处的光谱反射率,以及红边幅值和红边面积均可较好地表征水稻受稻纵卷叶螟危害后的卷叶率级别,可利用这两层次的光谱指标分别对小区水稻的受害程度进行诊断。
【Objective】 The objective of this study is to research the spectral characteristics from canopy, undamaged and damaged leaves of rice at the booting stage, and to build the regression models based on the spectral parameters to diagnose the damage levels of rice leaf folder (RLF) (Cnaphalocrocis medinalis). 【Method】 The reflectance from rice canopy, undamaged leaves and damaged leaves by RLF was measured by Hand-held Spectroradiometers, and correlation analysis was made to explore the sensitive wavebands of spectra to damage levels. The linear and stepwise regression methods were introduced to construct the diagnostic models of damage levels of rice based on the spectral parameters.【Result】Reflectance from rice canopy at the near infrared regions decreased significantly with the increase of damage levels of rice, and the 738-1 000 nm was the sensitive wavebands to exhibit the damage of RLF. The reflectance from the undamaged leaves collected from the damaged plot also had the capability to exhibit the damage levels of RLF, and there was a significantly negative correlation between the reflectance at 512-606 or 699-1 000 nm and the damage levels. However, the reflectance at 582-688 nm from the damaged leaves had a significantly positive correlation with the damage levels. Consequent change in amplitude and area of red-edge from canopy, undamaged and damaged leaves occurred significantly when the rice was damaged by RLF, and the regression models to diagnose the damage levels using the spectral parameters were built. The diagnosing error of the model based on the red-edge amplitude from rice canopy or based on the reflectance at 550 nm from the undamaged rice leaves was lower, and the stepwise regression model’s based on all the 21 spectral indices from canopy, undamaged and damaged leaves was the lowest. These models could be used to monitor the damage by RLF.【Conclusion】The reflectance at 738-1 000 nm from the canopy and at 512-606 and 699-1 000 nm from the undamaged leaves of rice could reflect well the damage levels of rice by RLF, and the spectral models based on reflectance and red-edge parameters from the canopy and undamaged leaves could be used to diagnose the damage of RLF.
出处
《中国农业科学》
CAS
CSCD
北大核心
2012年第24期5040-5048,共9页
Scientia Agricultura Sinica
基金
国家公益性行业(农业)科研专项(200903051)
关键词
水稻
稻纵卷叶螟
卷叶率
危害程度
光谱反射率
诊断模型
Oryza sativa
Cnaphalocrocis medinalis
leaf-roll rate
damage level
spectral reflectance
diagnosing model