基于LandSat8 OLI数据的山区阴影信息检测与提取
发布时间:2018-01-15 23:00
本文关键词:基于LandSat8 OLI数据的山区阴影信息检测与提取 出处:《山地学报》2017年04期 论文类型:期刊论文
更多相关文章: LandSat OLI影像 c!算法 阴影 坡度因子
【摘要】:为快速获取山区遥感影像上的阴影干扰区域,探究一种简便、高效、精确的遥感影像中阴影干扰区域的方法,具有重要的意义。以福建省长汀县为研究区域,搜集2016年3月美国陆地卫星影像数据与ASTER影像计算的GDEM V2产品。基于土地利用分类体系测量的6类地面物体光谱信息与影像波段信息,优选影像的光谱波段并重新组合;选取Sinh函数与Max函数建立c!算法,对LandSat8 OLI影像进行差异化计算,通过二分式判别规则初步提取阴影区域;加入由数字高程模型计算的坡度信息,剔除水域与坡度较为平缓的地物等干扰信息,精确提取山区阴影区域。设立网格随机设置精度验证点验证精度,最终总体精度达到99.06%,Kappa系数为0.98。结果表明,实验方法对于LandSat8 OLI影像提取阴影可行性高,检测效果与提取结果较c3算法与SVI指数更好。
[Abstract]:In order to quickly obtain shadow interference area in remote sensing image of mountain area, it is of great significance to explore a simple, efficient and accurate method of shadow interference area in remote sensing image. Changting County, Fujian Province is taken as the research area. Collecting American Landsat Image data from March 2016 and GDEM of ASTER Image Computation. V2 products. Spectral information and image band information of 6 kinds of ground objects measured based on land use classification system. Select the spectral band of the image and recombine it; Select Sinh function and Max function to establish c! The algorithm is used to calculate the difference of LandSat8 OLI image, and the shadow region is preliminarily extracted by dichotomous discriminant rule. Add the slope information calculated by the digital elevation model, eliminate the disturbance information such as water area and gentle slope, extract the shadow area of mountain area accurately, and set up the grid random setting accuracy verification point to verify the accuracy. The result shows that the experimental method is feasible for LandSat8 OLI image to extract shadow. The result of detection and extraction is better than that of c3 algorithm and SVI index.
【作者单位】: 福建农林大学林学院;
【基金】:生态林种科研基地建设工程项目(61201400814) 森林持续经营研究(ky0180081)~~
【分类号】:TP751
【正文快照】: 目前的遥感手段的基础主要源于地物对光照的反射,由于太阳高度角与地物高程差,势必容易产生对太阳光的阻挡从而产生一定区域的阴影覆盖范围。由于阴影的干扰,相同或者相似的地物类型在是否有阴影的不同环境下呈现较大的光谱、纹理特征差异,对后续的遥感分析造成了较大的干扰[1,
本文编号:1430420
本文链接:https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/1430420.html