陡峭山区影像的半经验地形校正

发布时间：2018-05-15 01:19

本文选题：地形校正 + SCEDIL校正　；参考：《遥感学报》2017年05期

【摘要】：本文提出一种新的半经验地形校正模型SCEDIL(Simple topographic Correction using Estimation of Diffuse Light),该模型通过结合DEM与光学影像数据寻找局部区域内完全光照和阴影的水平像元,并以光照、阴影水平像元的平均反射率值估算局部区域散射辐射比,提高了陡峭山区影像的地形校正精度。以高分一号卫星和Landsat ETM+影像为例,从目视判读和定量分析两个方面,比较分析该算法与传统半经验地形校正算法(C、SCS+C)的校正结果。结果表明:(1)对较为平坦的地形,SCEDIL和C、SCS+C校正都有较好的目视结果;对地面起伏较大的陡峭地形,C、SCS+C校正后,原阴影区域易呈现破碎化特征,SCEDIL校正后,原阴影区域过渡较为平滑。(2)SCEDIL校正后,各波段反射率的均值和标准差优于C、SCS+C校正,SCEDIL校正后,影像总分类精度与同类地物光谱信息均一性均优于C和SCS+C校正。SCEDIL半经验地形校正方法能有效地去除影像中的地形干扰,尤其对陡峭地形的校正效果,优于常规地形校正模型。
[Abstract]:In this paper, a new semi-empirical terrain correction model, SCEDIL(Simple topographic Correction using Estimation of Diffuse Lightning, is proposed. The model combines DEM with optical image data to find out the horizontal pixel of complete illumination and shadow in the local region, and to use illumination. The average reflectivity of the shadow horizontal pixel is used to estimate the local area scattering radiation ratio, which improves the terrain correction accuracy of the steep mountain image. Taking Gaofian 1 satellite and Landsat ETM images as examples, the correction results of this algorithm and the traditional semi-empirical terrain correction algorithm are compared and analyzed from two aspects of visual interpretation and quantitative analysis. The results show that: 1) there are good visual results for both SCEDIL and CSCS C correction of flat terrain, and after CSCS C correction for steep terrain with large surface fluctuation, the original shadow area is prone to present fragmentation characteristics and SCEDIL correction. The mean value and standard deviation of reflectivity of each band are better than those of CSCS C correction and SCEDIL correction after the smooth transition of original shadow region. The accuracy of image classification and spectral information of similar ground objects are both uniform, which is better than that of C and SCS C correction. SCEDIL semi-empirical terrain correction method can effectively remove terrain interference in image, especially for steep terrain, and is superior to the conventional terrain correction model.
【作者单位】：北京林业大学省部共建森林培育与保护教育部重点实验室;中国林业科学研究院资源信息研究所;
【基金】：国家重点基础研究发展计划(973计划)(编号:2013CB733401) 国家自然科学基金(编号:41571332) 北京高等学校青年英才计划(编号:YETP0741)~~
【分类号】：P237

【相似文献】