阴阳坡不同划分方法的坡度匹配模型地形校正研究

Topographic Correction of Sunny and Shady Slope in Different Division Methods Based on Slope-matching Model

基于坡度匹配模型,对香格里拉和高黎贡山两景Landsat5 TM影像,分别利用地理学、林业资源调查、阴影形成原理、成像太阳方位角和成像太阳天顶角5种不同阴阳坡划分方法求出相关参数对TM影像进行地形校正。结果表明:从视觉分析上看,校正后图像的立体感降低,在一定程度上消除了地形的影响。从阴阳坡均值上看,校正前阳坡和校正后阳坡均值相同,说明地形效应对阳坡几乎无影响;B划分校正结果各波段阴阳坡亮度均值差范围最小,表明B划分校正效果最好。从直方图上看,校正后减弱了地形效应,遥感图像的直方图近似正态分布;从直方图细节上比较,B划分校正结果最好。从变异系数上分析,校正后影像变异系数值相对校正前均有所减小,D划分在各波段上CV值最小,说明D划分校正结果最好。因此,经过地形校正能够消除影像大部分地形效应的影响;地形效应对阳坡亮度值影响不大,校正后的影像能较好地保持研究区地物在各波段上的光谱特性;比较5种阴阳坡划分方法地形校正结果,林业资源调查阴阳坡划分的校正效果最佳,其次为成像太阳天顶角、地理学、成像太阳方位角、地形阴影形成原理。

By using geography, forestry resource survey, shadow formation principle, solar azimuth of acquisi- tion time and solar zenith of acquisition time 5 different division methods of shade slope and sunny slope to calculate parameters for topographic correction of the Landsat5 TM images of Shangri-La and Gaoligong Mountain based on slope-matching model. From visual analysis, image perspectivethe reduced after correction, to some extent, to elim- inate the impact of the terrain. From the mean of shady slope and sunny slope, the same mean of sunny slope and sunny before correction and corrected to illustrate that the topography effect had little effect on the sunny slope. The range of mean brightness difference between sunny and shady slopes of each band was minimum in the correction results of B divide which was the best. From the histograms, topographic effects reduced after correction, histogram of remote sensing images approximated normal distribution. B was the best according to the comparison of the histo- gram details. Based on the analysis of coefficient of variation, coefficient of variation of corrected images was smaller than before correction, the D had the smallest CV value at each band, which showed that the D was the best. Most of the topographic effect was removed after terrain correcting on image, The topographic effect had little effect on the sunny slope brightness, overcorrection existed after correcting results by slope-matching model, but the spectral characteristics maintained better for each band features of study area on the corrected images. Comparison of the corrected results in 5 kinds of shady and sunny slopes division method, the forestry resource survey in divi- sion of shady slope and sunny slope was the best, followed by solar zenith of acquisition time, solar azimuth of ac- quisition time, geography and topographic shadow formation principle.