摘要
由DEM提取地形因子时一般采取3×3窗口分析运算,导致得到的地形因子矩阵在边缘上存在明显边缘误差。以黄土高原不同地貌类型区坡度的提取为例,研究基于1∶10 000和1∶50 000比例尺DEM提取坡度时所产生的边缘效应,分析了边缘效应与DEM分辨率、区域面积及区域地貌特征之间的关系,提出了解决或改善边缘效应的方法。用2个指标量化了边缘效应:一是平均坡度的变化值;二是边缘部分坡度提取的中误差。试验结果显示:平均坡度的变化值与DEM分辨率、区域面积及区域地貌特征之间有较明显的相关关系,即分辨率越高、区域面积越大,地面起伏越小,则边缘效应对整个区域平均坡度的影响就越小;而边缘部分坡度提取的中误差与区域面积没有明显的稳定相关关系,但与DEM分辨率和区域地貌特征之间存在较明显相关关系,即分辨率越高,地面起伏越大,则边缘上坡度提取的中误差越大。
Neighborhood analysis calculation with 3 × 3 window is usually adopted to derive slope from Grid DEM. This method produces errors on the edge of terrain factors matrix. This paper takes slope as an example, three areas with different terrain characteristics as test areas, studies the edge effect on deriving slope from DEM with different scale, analyses the relationship between the edge effect and DEM resolution, region area and terrain characters, puts forward some methods to solve or decrease impact of the edge effect. It is important for digital terrain analysis with high precision. There are two indexes to express edge effect： the first is the change value of mean slope; the second is root mean square error on the edge when deriving slope. The result shows that, the change value of mean slope is related to DEM resolution, region area and terrain characteristics, RMSE on the edge when deriving slope is related to DEM resolution and terrain characteristics.
出处
《水土保持通报》
CSCD
北大核心
2006年第3期82-85,116,共5页
Bulletin of Soil and Water Conservation
基金
国家自然科学基金(40271089)
