棉花黄萎病叶片氮素含量与高光谱数据相关性分析

The Relationship between Nitrogen Contents and Hyper Spectra Data in Leaves of Cotton under Verticillium wilt

【目的】探讨棉花黄萎病叶片全氮含量（LTNC）与高光谱数据间相关性,为遥感监测黄萎病冠层全氮含量提供理论依据。【方法】在棉花黄萎病发生的不同时期,获取黄萎病叶片光谱及氮素含量数据,对其光谱反射率进行一阶微分处理和光谱特征参数提取,并与氮素含量进行相关分析。【结果】（1）棉花黄萎病LTNC与反射光谱在404~728 nm,880~2 500 nm呈极显著负相关（平均r=-0.71,-0.68）,748~768 nm呈极显著正相关（平均r=0.31）,且在694、758、2 247 nm处的相关系数最大,｜r｜=0.87,0.32,0.73;（2）黄萎病LTNC与一阶微分光谱在700~753 nm处呈极显著正相关（平均r=0.89）,在452~632 nm处呈极显著负相关（平均r=-0.64）,与单波段554 nm处相关系最大（｜r｜=0.90）;（3）黄萎病LTNC与24个光谱特征参数均达极显著相关,组合波段参数相关性最好,植被指数和吸收参数的相关性较好,单波段相关性一般,与PVI[FD554,FD 731]的相关性最好,r值最大,为0.91;与PPR[550,450]的相关性最差,｜r｜仅为0.48。【结论】黄萎病棉叶LTNC与高光谱数据间有高的相关性,可利用高光谱数据对黄萎病LTNC进行遥感监测。

【Objective】The paper aims to provide a theoretical basis for monitoring cotton canopy nitrogen contents of Verticillium wilt by discussing correlation relation between hyper spectra data and cotton LTNC of the disease.【Method】The data of spectrum and leaves nitrogen contents（ LTNC） of cotton infected by Verticillium wilt were acquired in different cotton disease phases,and LTNC of the disease and spectra reflectance,the first derivative of reflectance and spectral characteristic parameters were extracted and went through the correlation analysis,respectively. 【Result】1. The correlations between LTNC and spectral reflectance were negative at the significance between 404- 728 nm,880- 2 500 nm（ mean r =- 0. 71,- 0. 68）,and positivesignificance between 748- 768 nm（ mean r = 0. 31）,the best significant single bands had 694 nm,758 nm,2 247 nm（ ｜ r ｜ = 0. 87,0. 32,0. 73）. 2. The correlations between LTNC and derivative spectra were positive best significant in 700- 753 nm（ mean r = 0. 89） and negative best significant in 452- 632 nm（ mean r =-0. 64）,the best significant single band was 554 nm（ ｜ r ｜ = 0. 90）. 3. There were best significant correlation between 24 spectral characteristic parameters and LTNC of cotton leaf infected disease. In the middle of them,Combination bands had best relationship,vegetation index and absorption parameter had better relationship,single bands had ordinal relationship with LTNC,and the PVI[FD554,FD731]had max positive correlation（ r = 0. 91）,the PPR[550,450] had mini negative correlation（ ｜ r ｜ = 0. 48）.【Conclusion】This study showed that there was a correlation relation between hyper spectra data and LTNC of disease,so it was viable monitor LTNC of the disease by using remote sensing.