基于RILBP的地形结构尺度稳定特征分析  

Scale Stability Investigation Based on RILBP for Terrain Structure

作　　者：李可[1] 晏实江[2] LI Ke;YAN Shi-jiang(School of Geograp hy Science,Nanjing Normal University , Nanjing 210023;School of Resources and Environmental Engineering ,Anhui University ,Hefei230601,China)

机构地区：[1]南京师范大学地理科学学院,江苏南京210023 [2]安徽大学资源与环境工程学院,安徽合肥230601

出　　处：《地理与地理信息科学》2017年第4期63-68,共6页Geography and Geo-Information Science

基　　金：国家自然科学基金项目(41601411;41671389;41401440);安徽大学2014年大学生科研训练计划项目(KYXL201404)

摘　　要：地形结构描述着地表起伏的形态结构,而高低起伏的地形结构又组成了丰富多样的地貌景观。地表形态一般呈现出局部形态相似与区域有序分异的尺度特征。现有的地形尺度研究不能有效地刻画地形结构的尺度特征,地貌形态的适宜表达尺度及地形结构的尺度稳定特征仍值得进一步探讨。该文以黄土高原不同地貌类型区为例,以数字高程模型(DEM)为数据源,从刻画地表形态的局部特征入手,系统地选取了多类具有代表性的地貌形态样本数据,分析了典型地貌类型区域的局部二值模式(LBP)分布特征。考虑到地形结构与RILBP之间一对多的对应关系,利用顾及旋转不变特性的LBP(RILBP)算法探究地形结构的尺度稳定特性,发现随着分析窗口增大,LBP显示出逐渐趋于稳定的尺度特征。研究结果表明,RILBP能够有效地表达地形结构,地形结构稳定单元能够有效地反映适宜的地形分析尺度,地形结构对地形多尺度信息的挖掘具有独特的优势。The terrain structure describes the morphological features of land surface.Rich and varied landscapes are attributed to the ups and downs of the terrain structure.Generally,the eart h surface morphology exhibit s different scale features of morpho-logic similarity at local scale and orderly differentiation at regional scale.Existing studies cannot effectively characterize the scale feature of terrain structure.T hus,the suitable expression scale of topography and stability features of morphological structure are needed to be further explored.1n this paper,a series of DEMs which belong t o different loess geomorphic types were selected as the source data.And Local Binary Pattern(LBP)was introduced to portray the spatial structure of central position and its neighborhood positions.This paper proceeds from p.)rtraying the characteristics of local surface configuration,and sequentially analyzes the distribution featm'e of LBP of typical loess landform types.A method of Rotation1nvariant LBP(R1LBP)was used to portray and analyze the landforms in consideration of the one to many relationships between topographic structure and LBPs.1n addition,the similarity between the scale samples was measured by calculating the Euclidean distance between theR1LBP fre-quency distribution vectors of the samples.1t is found that LBP tends tobe stablewith the increasing of the analysis scale.Final-ly,by using a minimum value among the intei-class difference of sample areas as a threshold to divide the Euclidean distance ma-trices,the stable cell sizes of all sample areas can be obtained.The results show that R1LBP can effectively express the terrain structure-Meanwhile,the stable unit of terrain structure can directly rflect the appropriate scale of terrain analysis and it has significant advantages in mining mult-scale topographical information.

关 键 词：局部二值模式 地形结构 稳定尺度 

分 类 号：P208[天文地球—地图制图学与地理信息工程]