文献类型： 外文期刊

作者： Song, Xiao ¹ ; Feng, Wei ¹ ; He, Li ¹ ; Xu, Duanyang; Zhang, Hai-Yan ¹ ; Li, Xiao ¹ ; Wang, Zhi-Jie; Coburn, Craig ¹ ;

作者机构： 1.Henan Agr Univ, Collaborat Innovat Ctr Henan Grain Crops, Natl Engn Res Ctr Wheat, 63 Nongye Rd, Zhengzhou 450002, Henan, Peoples R China

2.Henan Acad Agr Sci, Inst Plant Nutrient & Environm

关键词： Wheat;Viewing angle;Vegetation indices;Leaf N concentration;Monitoring model

期刊名称：ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING （ 影响因子：8.979； 五年影响因子：9.948 ）

ISSN：

年卷期：

页码：

收录情况： SCI

摘要： Real-time, nondestructive monitoring of crop nitrogen (N) status is a critical factor for precision N management during wheat production. Over a 3-year period, we analyzed different wheat cultivars grown under different experimental conditions in China and Canada and studied the effects of viewing angle on the relationships between various vegetation indices (VIs) and leaf nitrogen concentration (LNC) using hyperspectral data from 11 field experiments. The objective was to improve the prediction accuracy by minimizing the effects of viewing angle on LNC estimation to construct a novel vegetation index (VI) for use under different experimental conditions. We examined the stability of previously reported optimum VIs obtained from 13 traditional indices for estimating LNC at 13 viewing zenith angles (VZAs) in the solar principal plane (SPP). Backscattering direction showed better index performance than forward scattering direction. Red-edge VIs including modified normalized difference vegetation index (mND705), ratio index within the red edge region (RI-1dB) and normalized difference red edge index (NDRE) were highly correlated with LNC, as confirmed by high R-2 determination coefficients. However, these common VIs tended to saturation, as the relationships strongly depended on experimental conditions. To overcome the influence of VZA on VIs, the chlorophyll-and LNC-sensitive NDRE index was divided by the floating-position water band index (FWBI) to generate the integrated narrow-band vegetation index. The highest correlation between the novel NDRE/FWBI parameter and LNC (R-2 = 0.852) occurred at -10, while the lowest correlation (R-2 = 0.745) occurred at 60 degrees. NDRE/FWBI was more highly correlated with LNC than existing commonly used VIs at an identical viewing zenith angle. Upon further analysis of angle combinations, our novel VI exhibited the best performance, with the best prediction accuracy at 0 to 20 (R-2 = 0.838, RMSE = 0.360) and relatively good accuracy at 0 to 30 (R-2 = 0.835, RMSE = 0.366). As it is possible to monitor plant N status over a wide range of angles using portable spectrometers, viewing angles of as much as 0 to 30 are common. Consequently, we developed a united model across angles of 0 to 30 to reduce the effects of viewing angle on LNC prediction in wheat. The proposed combined NDRE/FWBI parameter, designated the wide-angle-adaptability nitrogen index (WANI), is superior for estimating LNC in wheat on a regional scale in China and Canada. (C) 2016 International Society for Photogrammetry and Remote Sensing, Inc. (ISPRS). Published by Elsevier B.V. All rights reserved.