基于DEM的河网提取算法的比较

Comparison of algorithms for extracting drainage network from grid-based digital elevation model

本文在数据独立的DEM解译算法分析方法支持下,以比汇水面积(Specific catchment area,SCA)为计算对象,对目前流行的5种路径算法D8、Rho8、FMFD、Dinf和DEMON进行了详细的比较和分析。文中首先从SCA的定义出发,建立了SCA的计算表达式,然后构建了表达式已知的4种数学曲面模拟DEM,最后通过SCA的计算值和理论值的统计分析,定量地描述了5种路径算法所具有的精度,并从算法复杂度、时间效率、健壮性、存储结构、格网结构及流向计算精度方面对路径算法进行了综合评价。结果认为基于坡向的路径算法即DEMON和Dinf能给出较可靠的SCA值和较理想的SCA分布,而基于坡度的路径算法的误差较小。

Based on the data-independent assessment of digital elevation model （DEM）, a quantitative analysis and comparison of five routing algorithm including D8, Rho8, FMFD, Dinf and DEMON, which are widely applied in GIS software and hydrological model is presented. The specific catchment area （SCA） is set as the calculated objective to be compared. The mathematical expression of SCA is established at first and then four mathematical surfaces including ellipsoid representing convex slopes, inverse ellipsoid representing concave slopes, saddle and plane are generated for computation of SCA by using these five routing algorithm. The computational results are compared with the theoretical value and the errors are statistically analyzed. The advantage and ~eakness of five algorithms are discussed. It is concluded that these routing algorithms based on slope direction such as DEMON and Dinf can provide reliable SCA and reasonable SCA distribution, but the routing algorithm based on slope gradient can provide the results with smaller errors.