利用几何代数的多维统一性与运动表达的简明性特征,设计了用几何代数向量场拟合进行地表坡向表达和提取的新算法。基于高斯合成曲面模型模拟的多尺度格网DEM数据,对坡向计算的正确性、抗噪性及空间整体性进行验证和统计分析,显示几何代数向量场拟合方法能正确提取地表坡向,且较之三阶反距离平方权算法具有更强的抗噪性;选取5类典型的不同地貌类型数据进行最优窗口和坡向提取试验,黄土丘陵,梁、峁、塬和风沙地貌的最优窗口大小分别为13、11、11、11、17,所提取的坡向可以较好地反映不同地貌类型的结构特征。
Based on the multidimension-unified and simple expression of movement characteristics of geometric algebra ,a new algorithm for expressing and extracting slope of surface multi-dimensional vector fields with the geometric algebra-based vector field matching method was proposed . The correctness , noise immunity and statistical spatial integrity of the slope extraction of our method was tested with multi-scale simulated grid DEMs generated with Gaussian composite surface . The results suggested the geometric algebra-based vector field matching method could correctly extract the slope of the surface ,and the algorithm had stronger noise immunity than three-order inverse distance square weight algorithm .The optimized window size finding and slope extraction experiments were constructed with five different real world data reflecting different terrain morphology . The optimized window sizes for loess hills ,loess ridge ,loess knolls ,loess plateau ,and loess sand geomorphology are 13 ,11 ,