摘要
为建立小麦株高的高光谱估算适宜模型,通过连续两年田间不同品种试验,在拔节至抽穗期同步测定小麦冠层高光谱数据和株高,并对两者的关系进行系统分析。结果表明,小麦株高与可见光波段呈负相关,与近红外波段呈正相关,与可见光波段的相关性总体上高于近红外波段。株高可以利用统一的光谱参数进行定量反演,其中以F698、D550、Dy、λr、SDr/SDb和(SDr-SDb)/(SDr+SDb)等光谱参数拟合效果较好。经两年的独立试验数据检验表明,以参数F698、D550、Dy及(SDr-SDb)/(SDr+SDb)为变量建立的株高估算模型表现较为稳定,尤其是以光谱参数(SDr-SDb)/(SDr+SDb)建立的模型,建模决定系数为0.85,预测决定系数和均方根偏差分别为0.86和4.27,相对误差为9%。因此,该参数可以作为估测小麦株高的有效光谱参数,对小麦生长中期的株高进行监测。
Preventing lodging in late growing stage and increasing crop yields are effective methods for enhancing income of farmers,but it depends on monitoring canopy reflectance spectra of plant height of winter wheat accurately and rapidly.Based on the field experiments with different cultivars during two growth years(2009-2010,2010-2011),canopy spectra data and measured plant height were collected from jointing stage to heading stage,systematic analysis was made for relationships between plant height and hyperspectral data.The results showed that the correlations between plant height and reflectance were negative in visible bands and positive in infrared bands,respectively.In general,the correlation between reflectance and plant height in visible bands was higher than in infrared bands.The regression analyses between canopy reflectance spectra and plant height revealed that some key spectral parameters could be used to estimate the changes in plant height across stages of growth with unified spectral parameters for each growth parameter.Among them,regression models based on F698,D550,Dy,λr,SDr/SDb and(SDr-SDb)/(SDr+SDb) were much better for plant height estimation.Testing of the monitoring models with an independent dataset indicated that the spectral indices of F698,D550,Dy and(SDr-SDb)/(SDr+SDb) could be used to estimate plant height in the growth season,of which(SDr-SDb)/(SDr+SDb) produced better estimation in model testing.Its decision coefficient of model was 0.85,the decision coefficient and the root mean square error of prediction were 0.86 and 4.27 separately,relative error was 9%.Hence,Plant height in winter wheat could be monitored by it with more reliable estimation in mid period of growth.
出处
《麦类作物学报》
CAS
CSCD
北大核心
2012年第3期523-529,共7页
Journal of Triticeae Crops
基金
河南省小麦产业体系项目(S2010-01-G04)
河南科技攻关项目(112102110030)
关键词
冬小麦
株高
高光谱
估算模型
Winter wheat
Plant height
Hyperspectral
Estimation models