冠层特征尺度的定量计算模型与方法

Canopy characteristic scale model and quantitative calculation

冠层特征尺度是植被定量遥感的基础概念,其物理定义和数学定量表达具有重要的研究意义。首先,基于光学辐射传输角度提出的冠层特征尺度的物理定义,即水平维线性混合条件下的最小分辨率单元,建立了冠层特征尺度的数学计算模型,并引入倒置的地统计学指数模型。然后,提出了基于局部方差分析的冠层特征尺度计算方法。最后,利用森林区域高分辨率图像,对论文提出的冠层特征尺度模型进行了定量验证。结果表明,冠层特征尺度模型计算的冠层特征尺度与树林株行距存在密切联系,线性复相关系数达0.95,证明了本文方法的合理性和可行性。本文提出的冠层特征尺度模型为地表特征尺度定量计算提供了一种新方法。

Canopy characteristic scale is a basic concept of quantitative remote sensing of vegetation, but also a very important feature, expression is of great significance for his physical definition and quantitative. The canopy characteristic scale physical connotation understanding and modeling mathematical expression, is the basic of research object linear and non-linear mixed, and it is the premise of optimal scale of observation objects.From the ray radiation transport point of view, there is a characteristic scale nonlinear mixed into the linear mixed transition, incident radiation between the transition characteristic scales of the object is inde- pendent of the optical properties, which can better describe the canopy group, choose the appropriate scale can get twice the result with half the effort to make remote sensing data, Base on the physical definition of optical radiation transmission proposed canopy characteristic scale, we established a mathe- matical calculating model for the canopy characteristic scale, and introducted the inverted geostatistics index model, put forward the calculation method of canopy characteristic scale based on analysis of local variance. Using the forest area of high resolution image, the canopy characteristic scale model proposed in this paper provide a quantitative validation. We chose 19 areas of higher density and vigorous growth forest as a study area, all the study areas with the same planting and regular distributed, and approxi- mately at the same stage of the operation. Data come from the Google Earth aerial photography data, with a resolution of 0.3 m, the data from two regions in Macon and Locke. The canopy characteristic scale model for analysis of data fi＇om the two regions, There is a certain correlation between the canopy characteristic scale model calculated and forest spacing measured, the linear correlation coefficient of 0.95, that showed spacing and vegetation canopy characteristic scale of the two regions is the presence of certain stable relationship, canopy charac- teristic scale calculation of canopy characteristic scale model was about 125 times the spacing. Canopy features scale models thesis, which can be in remote sensing images of forest vegetation growing season to choose their appropriate characteristic scale, canopy characteristic scale model is reasonable and feasible. Firstly, mixed pixel above the canopy characteristic scale belong to linear mixed, for its various transformations, you can use linear mixed pixel method to solve various scale difference or scale effect problems, Secondly, canopy characteristic scale is a basic object scale, when vegetation parameter population sampling （such as leaf area index）, canopy spectrum measurement, the sampling plots close to or greater than the canopy characteristic scale, can reflect the vegetation canopy component information; When the resolution is close to or below the canopy characteristic scale, ground surrounding pixel block or cross radiation effects to a minimum, arrive downward radiation level surface solar radiation spectrum and the corresponding pixel under the surface of the ground, ground reflectance can be calculated accurately.