Three-dimensionally consistent contour-based network rendered from digital terrain model data
Three-dimensionally consistent contour-based network rendered from digital terrain model data
复制标题
DOI：
10.1016/j.geomorph.2021.107969
复制
发表时间：
2021-09
期刊：
影响因子：
3.9
通讯作者：
H. Goto
中科院分区：
文献类型：
--
作者：
H. Goto
地理空间分析中常用的地貌学对象是基于等高线的网络（CBN）。与基于网格和补丁的网络不同，CBN垂直地划分了整个地形表面。传统的方法是根据正交性或最小距离度量将流线段与连续轮廓连接起来。这导致三维不一致（不是所有的直线段与轮廓正交相交）。因此，部分偏离了确切的流线。最陡下降法似乎是适用的；然而，它是为优化而不是跟踪而设计的。轨迹会在狭窄的丘陵和山谷中振荡，一旦遵循收敛准则，轨迹就会终止。无论是否接近一个精确的静止点，这种情况都会发生，但不存在延拓的方法。因此，作者设计了一个三维一致性CBN的框架。曲率矢量和公差有助于实现良好的再现性、存储效率和可调精度。该方法向精确曲线的内部推进，并在静止点上继续移动，坚持原始地形表面。通过包含另一种方法识别的轮廓，对流线曲线进行拾取，构成完整的CBN。最终的物体发出平滑的曲线，并具有三维一致性。利用低分辨率数字地形模型数据集进行了示范性实验；结果突出了所设计框架的重要性。
A common geomorphometric object in geospatial analysis is the contour-based network (CBN). Unlike grid- and patch-based networks, CBN partitions an entire terrain surface vertically. Traditional methods connect streamline segments with consecutive contours according to orthogonality or minimum-distance measure. This causes three-dimensional inconsistency (not all straight segments intersect orthogonally with a contour). Therefore, segments deviate from the exact streamline. The steepest descent methodology seems applicable; however, it was designed for optimization rather than tracking. A trajectory would oscillate at narrow hills and valleys, and tracing terminates once a convergence criterion follows. This occurs irrespective of proximity to an exact stationary point, but no method for continuation exists. Thus, the author devised a framework for a three-dimensionally consistent CBN. The curvature vector and tolerance help achieve good reproductivity, storage efficiency, and adjustable precision. The proposed method advances inward of the exact curve and continues moving over stationary points, adhering to the original terrain surface. A complete CBN is composed via gleaning the streamline curves by including contours identified by another method. The resulting object emanates smooth curves along with three-dimensional consistency. A demonstrative experiment is conducted using a lower-resolution digital terrain model dataset; the results highlight the significance of the devised framework.