An application of Coons patch to generate grid-based digital elevation models
An application of Coons patch to generate grid-based digital elevation models
复制标题
DOI：
10.1016/j.jag.2011.05.020
复制
发表时间：
2011-10-01
影响因子：
7.5
通讯作者：
Dai, Honglei
中科院分区：
文献类型：
--
作者：
Chen, Chuanfa;Li, Yanyan;Dai, Honglei
提出了一种基于Coons曲面片的格网数字高程模型（DEM）生成方法。采用一系列不同曲面复杂度的高斯合成曲面，对比分析了Coons曲面片与IDW、Natural Neighbor、SPLINE和KRIGING等经典插值方法的模拟精度。结果表明，在不考虑地形复杂性的情况下，Coons曲面片的精度比经典插值方法高得多，平均精度分别是IDW、Natural neighbor、SPLINE和KRIGING的327.30、15.19、3.66和40.07倍。一个真实世界的测试区域被用来测试采样密度对模拟精度的影响。结果表明，无论采样密度如何，Coons曲面片都比经典插值方法具有更高的精度。总之，Coons曲面片作为一种精确的插值方法，可以用于格网DEM的构建。(C)2011 Elsevier B.V.保留所有权利。
We have presented a new method for generating grid-based digital elevation models (DEMs) based on Coons patch. A series of Gaussian synthetic surfaces with various surface complexities were employed to comparatively analyze the simulation accuracies of Coons patch and the classical interpolation methods, such as IDW, Natural neighbor, SPLINE and KRIGING. Results indicate that irrespective of terrain complexity, Coons patch is much more accurate than the classical interpolation methods, and the accuracy of Coons patch is about 327.30, 15.19, 3.66 and 40.07 times as high as the accuracies of IDW, Natural neighbor, SPLINE and KRIGING on average. A real-world test area was used to test the effect of sampling density on the simulation accuracy. The results indicate that Coons patch has a higher accuracy than the classical interpolation methods regardless of sampling density. In a word, Coons patch as an accurate interpolation method can be used for grid-based DEM construction. (C) 2011 Elsevier B.V. All rights reserved.