我国典型林区的遥感影像地形校正方法效果评价.docVIP

  我国典型林区的遥感影像地形校正方法效 果评价# 王飒，陈晓玲，李熙，叶文韬** 5 10 15 20 25 30 35 40 （武汉大学测绘遥感信息工程国家重点实验室，武汉 430079） 摘要：本文比较了余弦校正模型、C 校正模型、SCS+C 校正模型、Minnaert 校正模型 4 个 经典模型在我国重点林区进行遥感影像地形校正的效果。江西和云南是我国重要的林区大 省，分别对应丘陵和山区地貌，对于遥感影像地形校正具有较强的研究意义。本研究采用两 省的 Landsat 5 遥感影像作为研究对象，实现了上述四种典型的地形校正算法，并从视觉角 度、影像统计角度和光照系数与像元值的相关性三个方面对地形校正方法进行定性和定量评 价。经过比较分析后发现，对于江西影像数据，C 校正方法是最理想的，余弦校正方法是不 适宜的，易出现校正过度现象；对于云南省影像，SCS+C 校正方法是最优的，余弦校正方 法同样不适用。 关键词：遥感；地形辐射校正；林区；TM 影像 中图分类号：TP751 Evaluating topographic correction methods of remote sensing images for typical forest areas in China WANG Sa, CHEN Xiaoling, LI Xi, YE Wentao (State Key Laboratory of Information Engineering in Surveying, Mapping and RemoteSensing, Wuhan University, WuHan 430079) Abstract: Compared the topographic effects in remote sensing images for typical forest in China with the four classical methods:Cosine correction model,C correction model,SCS + C correction model, Minnaert correction model in this paper.Jiangxi and Yunnan,they are the very important forest areas in China ,and hills are the normal landforms in Jiangxi,the mountainous area is the main in Yunnan,so they have the strong research significance in remote sensing images for topographic correction.Therefore,we use Jiangxi and Yunnan province as representatives of the forest areas in China ,and do the topographic correction experiments in Landsat 5 images.Finally, from the aspects of visual differcences,the imagesquantitative statistical and the relevance between the illumination coefficient and the pixel values,validated the four models,given the qualitative and quantitative evaluation.After analysis and comparison,for Jiangxi image data, the C-correction method has good correction performance and cosine correction method is most inappropriate for correction excessive phenomenon; image data for Yunnan Province, the SCS + C correction method is optimal, the cosine correction method is also bad to apply. Keywords: Remote sensing; Terrain radiometric correction; Forest areas; T