黄土高原丘坡信息DEM提取算法的应用

The Application of Algorithms for Extracting Slopes in Hilly Area of Loess Plateau Based on DEM

选择代表黄土地貌类型(丘陵沟壑区、梁峁区、高原丘陵过渡区)的75个样区(每个样区约4km2),以1∶1万水平分辨率为5m的DEM为研究对象,研究不同算法对提取地面坡度精度的影响。采用6种不同算法分别提取坡度。运用方差分析,对比分析,排序分析等方法,通过比较不同算法所提取坡度矩阵的平均值、最大值、标准差,以及坡度中误差等指标,认为在黄土丘陵沟壑区3类地貌类型区域如果整体考虑坡度的以上指标时选用三阶反距离平方权差分算法较为合理。在黄土高原沟壑区与丘陵沟壑区交错过渡地带也可以选用算法4提取坡度。同时,根据水土保持工作实际生产部门需要,提出并实现了对坡度信息评价时须分级别进行的思路。研究认为,在黄土丘陵沟壑区域相同的地貌类型区,算法对所获得的相同的坡度级别面积影响是不同的,并且规律不相同。在前述三种地貌类型区域,如果需要获得坡度分级面积信息,使用三阶反距离平方权差分算法和三阶反距离权差分最为理想。由于简单差分算法小于其他算法提取坡度所涉及的区域单元大小,不适宜在本区域使用。

In the 75 sample areas （4 km^2 each） selected from three regions with different physiognomic types in loess hilly-gully area, northern Shaannxi province, DEM with a spatial resolution of 5m according to the map scaled at 1：10000 is taken as the research object and 6 algorithms for deriving slopes are adopted. Employing the variance analysis, the author compares the slopes derived by different algorithms in such aspects as the mean, the maximum value, the standard deviation, the median error, and then concludes that if the above indexes are taken into consideration together, the reasonable algorithms are those of three-order inverse distance square weight difference and three-order inverse distance weight difference for the three regions with different physiognomic types in the loess hillly-gully area. Since the unit area involved in deriving slopes by difference calculation is smaller than that by other algorithms, the former is inappropriate to be used in this region. According to the actual requirement of production, the concept that the slope should be studied by dividing it into different grades is proposed. The study indicates that in the regions with the same physiognomic types in the loess hilly-gully area, the influences of algorithms on the slope area of the same grade are different, and the rules are different too. Algorithms should be selected according to the physiognomic types and the required grade of slope area in conserving water and soil. The paper can be referred to as concrete standards for selection.