论文部分内容阅读
地形纹理是区分不同地貌形态的重要依据,DEM是地形纹理分析的重要数据。然而,DEM分辨率使地形纹理特征提取存在着不确定性问题。本文以具有显著地貌多样性与差异性的陕西省为例,选择6个不同地貌类型区为研究区,以25 m分辨率DEM数据作为信息源,构建了多尺度的地面坡度、光照模拟和粗糙度数据序列。在此基础上,引入空间灰度共生矩阵(GLCM)对地形表面纹理特征进行量化分析,以揭示数据分辨率对地形纹理特征提取的影响。研究表明:对于单一样区,在DEM及其3个派生数据中,原始高程数据和粗糙度数据的纹理参数特征值,对分辨率的变化较为敏感。对于不同的地貌类型区,二阶角矩和对比度这2个纹理参数具有最大的变异系数,表明它们对于区分不同地貌类型的能力最强;二阶角矩具有较大的尺度依赖性,随着分辨率的降低,其区分能力急剧降低,而对比度对于地貌的区分能力,则随着分辨率的降低而增强,并保持在一个较大的范围内。DEM数据的对比度对于不同地貌的区分能力,在所选4个参数中最为稳定,而粗糙度数据的二阶角矩区分不同地貌的能力,随着数据分辨率的变化而最不稳定。以上结果对于根据不同的研究对象选择适宜的DEM分辨率及地形纹理参数具有一定的指导意义。
Topographic texture is an important basis for distinguishing different topographic features. DEM is an important data for topographic texture analysis. However, the DEM resolution has the problem of uncertainty in topographic texture feature extraction. Taking Shaanxi Province with obvious landform diversity and diversity as an example, this paper selected 6 different landform types as the study area, and built 25 m DEM data as the information source to construct multi-scale ground slope, light simulation and roughness Degree data sequence. On this basis, the spatial gray level co-occurrence matrix (GLCM) is introduced to quantitatively analyze the topographic surface texture features to reveal the effect of data resolution on the topographic texture feature extraction. The results show that for the single sample area, the texture parameters of the original elevation data and the roughness data are more sensitive to the change of resolution in the DEM and its three derived data. For different types of landforms, the second-order moment and contrast have the largest coefficients of variation, indicating that they have the strongest ability to distinguish different types of topography. The second-order moments have a larger scale dependence The resolution is reduced, its ability to differentiate sharply reduced, while the contrast for the ability to distinguish the topography, with the resolution decreases, and to maintain a large range. The contrast of DEM data is the most stable among the four selected parameters for the discrimination of different landforms while the second-order moment of roughness data distinguishes the ability of different topographic features and is most unstable with the change of data resolution. The above results have some guiding significance for choosing suitable DEM resolution and topographic texture parameters according to different research objects.