Efficient computations of geodesic distance
(English)Manuscript (preprint) (Other academic)
We present a novel way to efficiently compute anisotropic distances over a tessellated domain in two dimensions. The method is based on an integral formulation of distance and entails solving a dynamic programming problem. We also present an intuitive geometric construction that is used to characterize dierent types of boundary conditions and show how they aect the resulting distance function in our and competing work.
The included benchmark study shows that our method provides signicantly better results in anisotropic regions and is faster than a current stat-of-the-art solver. Additionally, our method is straightforward to code; the test implementation is less than 150 lines of C++ code.
Distance map, Geodesic distance, Riemannian manifolds
Engineering and Technology
IdentifiersURN: urn:nbn:se:liu:diva-54828OAI: oai:DiVA.org:liu-54828DiVA: diva2:310615