Abstract:Geological reconstruction is an intractable problem due to the noisy sample point set, the unevenness of sampling as well as the complicate topology. This paper presented a geological reconstruction algorithm based on Chaine's method, which perform a geometric convection process after the Delaunay tetrahedronization has been constructed. The major contribution of this paper included: Firstly, we adjust the shrinking order to make the result mesh as smooth as possible. Secondly, we introduce a topology cleanup process which could eliminate the facets that are not manifold. Finally, when dealing with surface like reconstruction, we force the shrinking process keep going until all output facets are thin part facets. Compared with other methods, our method runs fast and utilizes the original sample points as the vertices of the output mesh which is very useful in geological reconstruction. Experimenting with publicly available scan data demonstrates that our reconstruction algorithm is effective and efficiency. 1548-7741/Copyright © 2008 Binary Information Press October 2008.

Fast geological surface reconstruction from noisy point set