高光谱遥感特征参数与棉花产量及其构成因子的关系研究  

Relationship between Cotton Yield Components and Their Hyperspectral Remote Sensing Characteristics

作　　者：侯新杰[1] 蒋桂英[1] 白丽[2] 王冀川[1] 凌红波[1] 季春辉[1] 

机构地区：[1]石河子大学农学院 [2]石河子大学生命科学学院,石河子832003

出　　处：《遥感信息》2008年第2期10-16,共7页Remote Sensing Information

基　　金：新疆石河子大学引进博士基金项目,(505-822111)

摘　　要：通过大田和室内实验,测定了2个品种、4种梯度水、肥处理的棉花产量形成的关键时期———盛花期至吐絮期的冠层光谱反射率和产量及构成因素,对高光谱特征参数与产量构成因素进行相关统计分析。结果表明,棉花产量与抗大气植被指数VARI_700的相关性为0.9564;棉花产量构成因素中的单位面积总铃数、单铃重与棉花冠层光谱反射率之间相关性达到极显著水平;建立光谱特征参数与产量构成因子的回归方程,表明运用抗大气植被指数VARI_700和光谱曲线反射峰参数P_Depth554来反演单位面积总铃数是可行的,用光谱植被指数[820nm,1650nm]可估算棉花单铃重。The hyperspectral reflectances of the canopy of 2 sorts of cotton were measured by a ASD Field Spec Pro FRTM in field and indoor under 4 nitrogen and 4 irrigation application levels in growth process. At the same time, get date of the yield component trait ,for example, cotton bolls number and cotton bolls heavy per plant. Analysing the statistical correlation between characteristic variables of hyperspectral and yield component. The results showed that the better correlation between yield component and characteristic variables of hyperspectral than yield . The statistical results show that yield component is positively correlative with spectral reflectance in NIR bands , negatively correlative in visible light bands and shortwave-infrared bands for four growth stages. The normalized difference spectral indices combined by [560nm, 670nm], [670nm, 890nm], [890nm, 980nm], [920nm, 980nm], [857nm, 1210nm], [820nm, 1650nm], [820nm, 2200nm],[P_Depth554] are significantly correlative with yield component in growth stages, respectively. As for cotton bolls number ,VARI_700,P_Depth554 is especial significantly correlative with it in four growth stages. The [820nm, 1650nm] is positive correlation with cotton bolls heavy in four stages. So we can bulid models by VARI_700, P_Depth554, [820nm, 1650nm] for monitoring cotton growing trend and forecasting yield.

关 键 词：高光谱遥感 特征参数 棉花产量 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]