作　　者： ; ; ; ; ; ;

机构地区： 中国农业科学院农业资源与农业区划研究所农业部植物营养与养分循环重点开放实验室

出　　处： 《麦类作物学报》 2010年第1期96-100,共5页

摘　　要： 为避免土壤背景对冠层光谱的干扰,提高冬小麦定量化估产精度,以河北廊坊中低产田条件下的冬小麦为研究对象,利用ASD Field Spec高光谱仪定点获取冬小麦冠层光谱信息,分析了田间植被覆盖度和冠层NDVI在生育期内的变化,并利用植被覆盖度对冠层NDVI进行了校正。结果表明,通过三基色即RGB、色度和亮度可将数字图像中冬小麦和土壤背景进行分割,从而获得单位面积上冬小麦的覆盖百分比。而通过覆盖度校正后的植被指数即C-NDVI能够更具针对性地反映植株冠层氮素信息。在本试验条件下利用灌浆中期的C-NDVI与产量进行一元回归或利用全生育期的C-NDVI与产量进行多元回归均取得了较好的效果,决定系数分别为0.849和0.853。由于多元回归模型考虑了不同时期的C-NDVI的变化,因此模型具有更强的可靠性和稳定性,较适合于冬小麦定量化估产。 To avoid interference of soil on canopy spectrum and improve the estimation accuracy of wheat yield, in this experiment, based on winter wheat varieties （Triticum aestivum L. ） under the condition of middle-low yield fields, canopy reflectance were monitored using ASD FieldSpec Pro spectrometer （Spectral range from 350 nm to 2 500 nm） and the monitoring was conducted in the fixed site in the whole growing stages. The change trend of vegetation cover was analyzed, NDVI and C-NDVI were calculated from NDVI and vegetation cover within growing stages. The results showed that wheat can be separated from the soil based on their differences in RGB, H, and L values, so the percentage of wheat in unit area can be obtained. C-NDVI is the NDVI modified by vegetation coverage, which increased pertinence to spectral information of wheat canopy. In this paper, good prediction results were obtained from single regression of yield and C-NDVI in grain filling stage or multiple regression of yield and C NDVI in different stages, and the decision coefficients were 0. 849 and 0. 853, respectively. Comparatively, the multiple regression model is more reliable and stable because the changes of C-NDVI in different stages were considered in the model This method greatly improves the wheat yield estimation accuracy using ground spectral data.

关 键 词： 冬小麦 植被覆盖度 高光谱 估产

领　　域： [农业科学] [农业科学] [农业科学]