Simplification 3d Points Cloud Method Based On Importance Of 3d Points

Abstract

Representing the surface of complex objects, the samples resulting from their digitization can contain a very large number of points, endorsing simplification techniques, which analyse the relevance of the data. Simplification techniques also provide models with fewer points than the original ones. On the contrary the reconstruction of a surface, with simplified point cloud, must be close to the original. In this article, we develop a method of simplification reduce the number of scanned dense points, based on the density estimation and the notion of entropy. The performance of this approach is illustrated through experimental results and comparison with other simplification methods.