中文摘要：

以黄萎病胁迫下棉花叶片为试验材料, 分析感染病害棉叶叶绿素（Chl）含量和氮素含量（LNC）与光谱红边参数间的关系, 建立病害棉叶Chl含量和LNC的光谱红边参数诊断模型。结果表明：（1）随着病情加重, 棉叶Chl a、Chl b、Chl a＋b和LNC逐渐减小, 其中Chl a下降最快, Chl b下降最慢；（2）黄萎病叶片光谱反射率在可见光区（400~700 nm）, 近红外光区（700~1300 nm）和短波红外光区（1300~2500 nm）呈现逐渐上升趋势, 在520~680 nm间达极显著（P〈0.01）；光谱吸收率在可见光区和短波红外光区呈现逐渐下降的趋势, 达极显著（P〈0.01）, 在近红外光区呈现先升后降的趋势。（3）病害棉叶红边位置（REP）、红边振动幅（Dr）、红谷位置（Lo）、红边深度（Depth672）和红边面积（Area672）的值均减小, 红边宽度（Lwidth）的值增加, 且Area672减小的幅度最大, Dr减小的幅度最小, Lwidth增加的幅度较大；（4）病害棉叶Chl a含量、Chl b含量、Chl a＋b含量和LNC均与红边参数REP、Lo、Depth672和Area672呈极显著正相关, 与Lwidth呈极显著负相关, 与Dr未达显著相关；（5）利用红边参数建立的棉叶Chl含量和LNC的诊断模型均达极显著（P〈0.01）, 其中以Area672为自变量建立的病害棉叶Chl a、Chl a＋b和LNC的诊断模型和Lo为自变量建立的Chl b诊断模型的精度最高, 能很好的诊断病害棉叶Chl含量和LNC。

英文摘要：

The relationship between chlorophyll （Chl） content, leaf nitrogen content （LNC） and red edge parameters were analyzed, and diagnose models of spectra red edge parameters were established for cotton leaf infected by Verticillium wilt. Results showed that： （1） Chl a, Chl b, Chl a＋b, and LNC decreased with increasing in severity level （SL） of Verticillium wilt in cotton leaves, in which Chl a showed the highest and Chl b showed the lowest decrement rate, respectively. （2） Spectrum reflectance increased with increasing severity of Verticillium wilt in the visible region （400–700 nm）, near-infrared region （700–1300 nm） and short infrared region （1300–2500 nm）, and significantly higher increment was detected in 525–680 nm region （P〈0.01）. Spectrum absorption decreased significantly with increasing SL of Verticillium wilt in the visible region and short infrared region （P〈0.01）, and which increased first and then decreased in near-infrared region. （3） Decrease of REP, Dr, Lo, Depth672, Area672 and increase of Lwidth was detected among red edge parameters, in which Area672 showed the highest and Dr showed the Lowest decrement rate, respectively. （4） There was significant positive correlation between Chl a, Chl b, Chl a＋b, LNC of cotton leaves and REP, Lo, Depth672, Area672 of red edge parameters, significant negative correlation was found for Lwidth of red edge parameter, while no significant correlation was found for Dr of red edge parameter. （5） Diagnose models of Chl a, Chl a＋b, and LNC for Verticillium wilt in cotton leaves with the independent variables Area672, and Chl b with the independent variables Lo reached the best estimated precision （P〈0.01）. This could diagnose severity level of Verticillium wilt in cotton leaves effectively.