一种基于GeoEye-1立体像对数据的快速提取平均树高方法

A Method of Mean Forest Height Fast Extraction from GeoEye-1 Stereo Pairs Data

探讨了可见光立体像对遥感数据在森林平均树高估算研究方向的可行性,为解决大区域快速提取森林平均树高参数的科学问题提供技术支撑。利用GeoEye-1卫星立体像对中提供的有理多项式系数(RPC)参数和数字表面模型(DSM)与数字高程模型(DEM)的理论原理,建立了基于DSM和DEM空间相差模型建立林分冠层高度估算方法流程。结果表明:基于湖南攸县黄丰桥国有林场GeoEye-1立体像对影像数据,按照估算流程,最终得到试验区小班尺度的样地平均树高遥感提取结果。结合样地地面实测控制点和地面小班数据调查数据,该方法提取的研究区平均树高总体误差率在83.1%,其中最大误差为3.773 m,最小误差为0.025 m。因此,本研究是一种可以快速获得研究区大范围森林平均树高参数的创新、可行的方法。

：This paper discusses the feasibility of the remote sensing data to estimate the average tree height of forest, and provides the technical support for solving the problem of extracting the average tree height parameters of the forest in a large area. The processing method of Mean Forest Height extraction are researched based on mathematical model of the spatial difference between DSM and DEM. The canopy height model （CHM） of GeoEye-1 stereo pair data using rational function model was analyzed and experiment was carried out by using stereo pair of state-owned forest farms of Huangfengqiao, Hunan Province. In case with high accurate ground control points and ground data survey, the results show that the mean forest height accuracy of GeoEye-lstereo image pair is 83.1% （maximum error is 3.773 m and Minimum error is 0.025 m）. It is proved that stereo-pair data extraction method is an innovative and reliable method and workflow to obtain the mean height parameter of forest in large scale area.