基于最大熵模型的玉米冠层LAI升尺度方法

发布时间：2018-02-26 19:24

  本文关键词： 遥感 作物 算法 最大墒模型 叶面积指数 升尺度 尺度效应 环境变量　出处：《农业工程学报》2016年07期 　论文类型：期刊论文

【摘要】：叶面积指数(leaf area index,LAI)是表达农作物冠层结构的关键参数之一,准确获取LAI对于农作物长势监测、估产等研究具有非常重要的意义。由于地物空间复杂性、数据源的不同以及遥感反演模型的非线性,LAI的反演结果会存在尺度效应,因此需要进行尺度转换研究。理想的升尺度转换应该只是数据空间分辨率的降低,而数据内在信息应保存到低分辨率中。最大熵(maximum entropy,Max Ent)模型是基于多种环境因子的广义学习模型,对分析因子的空间分布具有较高的估算精度,因此,该研究利用最大熵模型进行玉米冠层LAI升尺度方法研究,从而将野外实测的LAI点数据扩展到空间分辨率为30 m的面数据,所使用的数据源是Landsat8 OLI遥感影像、气象数据和野外样点上测量的LAI数据。研究结果表明:利用最大熵模型升尺度转换结果与实测LAI相比,R2为0.601、RMSE为0.898,说明两者的相关性较高;由于玉米冠层叶片之间的相互遮挡,导致整体结果偏低,但偏低误差在可接受范围内。因此,Max Ent模型可用于农作物LAI点数据到面数据的升尺度转换。
[Abstract]:Leaf area index (Lai) is one of the key parameters to express the canopy structure of crops. The accurate acquisition of LAI is very important for the study of crop growth monitoring and yield estimation. The different data sources and the nonlinear inversion results of remote sensing inversion model will have scale effect, so scale conversion should be studied. The ideal scaling conversion should only reduce the spatial resolution of data. The maximum entropy maximum entropyMax Ent model is a generalized learning model based on a variety of environmental factors, which has a high estimation accuracy for the spatial distribution of analysis factors. In this study, the maximum entropy model was used to study the LAI scaling method of maize canopy, and the field measured LAI data was extended to the surface data with spatial resolution of 30 m. The data source was Landsat8 OLI remote sensing image. Meteorological data and LAI data measured on field samples. The results show that compared with the measured LAI, the RMSE of the maximum entropy model is 0.601g, which indicates that the correlation between them is relatively high, because of the mutual occlusion between maize canopy leaves, As a result, the overall result is low, but the error is within acceptable range. Therefore, Max Ent model can be used to transform crop LAI point data to surface data.
【作者单位】： 中国农业大学信息与电气工程学院;黑龙江省农垦科学院科技情报研究所;
【基金】：国家自然基金项目(41371327)
【分类号】：S513;S127
，

本文编号：1539348