用几何光学模型对方向变化和直方图建模(英文)

Modelling Directional Variance and Variograms Using Geo-Optical Models

森林和林地的图象二向性反射函数是一个统计函数，较之小尺度的树冠，它更多地用于大尺度的均匀覆盖的地块。用航空象片和高空间分辨率扫描仪数据作的图象热点影响研究在小尺度下显示出很大方差，而且，太阳和观测角度的交互变化进一步增加了这种反射各向异性变化的方差并有规律地继续呈现在分辨率低的图象中，这被称作ＢＲＶＦ或二向性反射方差函数。近年来，作为一种解释结构的手段，高分辨率图象的方向性方差和直方图结构越来越受到重视，这方面的数据也越来越多。这项工作是利用图象方差来解释结构问题（Ｓｔｒａｈｌｅｒ和李小文倡导）的一种延伸，并在过去１５年中由众多人员作了大量工作。在树冠尺度下，森林的直方图和二向反射方差函数可以计算出来，这里利用了近似迭代函数来处理这些数据并和数值积分模拟进行了对比，结果显示可对二向性反射方差函数的测量和引入直方图的各向异性进行准确建模。

The image BRDF of forests and woodlands is a statistical function which operates at the scale of an average patch of cover rather than at the scale of crowns. Studies of the image Hotspot effect using aerial photography and high spatial resolution scanner data shows very high variance at this detailed scale. In addition, the directional effects of the sun and observer positions interact significantly to create an angular anisotropic variation which persists up to aggregated scales. This has been called the BRVF or Bidirectional Reflectance Variance Function. There has been a recent growing interest in the directional variance or variogram structure of high resolution images as a means to interpret structure and such data have become regularly flown. This work is an extension of the use of image variance to interpret structure as pioneered by Strahler and Li and explored by various authors over the last 15 years. Directional variograms and BRVF functions for forests at the crown scale can be computed using approximations to the overlap functions driving these second order statistics and compared with numerically integrated simulations. It has been shown that the scaling of BRVF and the anisotropy introduces into the variogram can be accurately modelled.