Spherical space limited triangle triangulation

Question

In order to implement a high-performance dynamic algorithm for tracking trajectories on the sphere (in C ++), I am interested in performing incremental limited delaunay triangulation on the surface of the sphere. Existing libraries seem to be inadequate - the closest to what I could find so far is CGAL, which has a topological space on the right, but the metric space is wrong.

The library must have:

• Reasonable performance (I have about 100 thousand points for input)
• Spherical topological and metric space (to be honest, this overlaps # 1 by a large margin)
• Insert and delete incremental points (for subsequent use of algorithms)

At the moment, my only real parameters seem approximate (using projection onto a two-dimensional Euclidean metric space and taking compromises in the Delaunay guarantee that it provides) or write my own, with all the probability that it entails. Is there a library for delaunay limited triangulation in spherical metric space?

Answer 1

Since this is now flagged as irrelevant (due to spam?), I could also give at least an exhaustive answer.

As far as I know, as of July 2017, there are two options for calculating the limited triangulation of a triangle of points on a sphere.

- libdts2 (1). 2D CGAL 2D , . , , , (2). 4 , . , . , .

- (3). . , , CGAL (4).

, libdts2 , GeometryFactory/INRIA . , libdts2 CGAL. ​​

libdts2 , :

• CDT OpenStreetMap Saarland (309K , 324K ) 16 , 170 MiB Ram ( , Core i7-4700MQ)
• ,
• / , CGAL

:

• libdts2 GitHub
• " : "
• " Delaunay
• CGAL Google Summer of Code