| 齐雁冰,楚万林,解飞,陈洋,常庆瑞.基于高光谱的渭北旱塬区棉花冠层叶面积指数估算[J].干旱地区农业研究,2017,35(1):114~121 |
| 基于高光谱的渭北旱塬区棉花冠层叶面积指数估算 |
| Estimating leaf area index of cotton canopy by hyperspectral reflectance in Weibei plateau |
| |
| DOI：10.7606/j.issn.1000-7601.2017.01.18 |
| 中文关键词: 棉花 叶面积指数 高光谱 植被指数 |
| 英文关键词:cotton leaf area index hyperspectrum vegetation index |
| 基金项目:“十二五”农村领域国家科技计划课题（2013AA102401-2） |
|
| 摘要点击次数: 1667 |
| 全文下载次数: 787 |
| 中文摘要: |
| 以棉花冠层高光谱反射率与冠层叶片叶面积指数(LAI)为数据源，在分析LAI与原始高光谱反射率、一阶微分光谱反射率、光谱提取变量和植被指数相关性的基础上，采用一元线性与多元回归的方法构建了棉花LAI高光谱估算模型，并进行精度估算。结果显示，在可见光范围内随着生育期的推进及施氮量的增加冠层光谱反射率逐渐降低，在近红外范围内从苗期到花铃期随着施氮量增加反射率逐渐增加，花铃期到吐絮期反射率明显降低；各生育期冠层光谱的提取变量与LAI的相关性不强，全生育期各种光谱提取量及植被指数与LAI的相关性高于不同生育期；棉花冠层叶片LAI在反射光谱1 461 nm处相关系数达到最大值(r=-0.726)；对于一阶微分光谱，LAI的敏感波段发生在742 nm处，r=0.744；以敏感波段742 nm一阶微分光谱反射率建立的逐步回归估算模型精度最高，RMSE=0.94，RE=26.27%，r=0.78。说明以全生育期为基础，采用一阶微分光谱敏感波段，并根据实际条件选择有效的估测模型，可以进行棉花LAI的预测。 |
| 英文摘要: |
| Leaf Area Index (LAI) is an important parameter to assess the growth situation of cotton. In this paper, canopy hyperspectral reflectance and LAI were measured at six growth stages of cotton in a field experiment. The correlation of LAI with the original spectral reflectance, the first derivative spectral reflectance, commonly used spectral variables and vegetation index were analyzed. The estimation models of LAI were established using linear regression and multiply stepwise regression methods, and the predictive precision was analyzed. The results indicated that spectral reflectance of cotton canopy decreased gradually with the advance of the growth stage and increase of nitrogen fertilizer application in the visible band, while it was increased from the seeding stage to the flowering and ball stage and it was decreased from the flowering and ball stage to the ball opening stage in the infrared band. The correlation coefficient of LAI with the common used spectral variables and vegetation index were higher in the whole growth stage than the different stages. The maximum correlation coefficients of LAI occurred at the reflectance bands of 1 461 nm with the r=-0.726, while the highest correlation coefficients between the first derivative spectral data and LAI occurred at band of 742 nm with r=0.744. The model based on the first derivative spectral reflectance by using multiply stepwise regression method obtained the most satisfied results for the estimation of LAI in the 742 nm, RMSE=0.94, RE=26.27%, r=0.78. It is feasible to monitor the cotton growth by the first derivative spectral reflectance based on data of the whole growth stage. But for the different regions, the estimating models should be assessed carefully based on plenty of experiments. |
| 查看全文 查看/发表评论 下载PDF阅读器 |