The Ridge-River network is the union of the drainage network and the ridge network. Our research has shown that these points are the most significant for compressing the terrain while minimizing the amount of hydrology-related error into the terrain reconstruction. The ridge-river network is refined using the Douglas-Peucker algorithm to extract the most significant points. Algorithms for compressing and reconstructing the terrain based on these points are cited in the publications section as well as one this site.
]]>
A new data structure for simplifying terrain that captures hydrology significant features using a constrained Delaunay triangulation. This constrained triangulation preserves the hydrology by using irregular-sized, non-overlapping planes to model regions that flow in a uniform direction. Constrained edges are associated with drainage and ridge networks that incorporate physically-based structure into the model without significant overhead. This allows better compression ratios then standard Triangulated Irregular Networks with higher accuracy. In particular, hydrology significant error is reduced. Standard error metrics such as root mean squared (RMS) and maximum error fail to capture whether a reconstructed terrain accurately captures the hydrology. A hydrology error metric is used to verify our results based on the potential energy required for the reconstructed drainage to flow on the original terrain.
]]>
A quantitative measurement of how well a drainage network captures the hydrology is very important for determining how well terrain simplification technique preserves the hydrology. Past work has been done for comparing how well a computed drainage compares to the real world drainage. Most of the time, real world flow measurements are unavailable and flow simulations have to be used to make predictions (floods, erosion, pollutants, etc). Having a measurement for testing and comparing different models has the potential to be widely used and practical across numerous applications.
]]>