Error-Controllable Simplification of Point Cloud

Point cloud simplification has become a vital step in any point-based surface processing pipeline. This paper describes a fast and effective algorithm for point cloud simplification with feature preservation. First, feature points are extracted by thresholding curvatures; Second, for non-feature points, they are covered by distinct balls, the points in each ball are substituted by an optimized point. Thus, the simplified point cloud consists of extracted feature points and optimized points. This algorithm is able to produce coarse-to-fine models by controlling a general error level. But the error level of each ball may be adaptively adjusted according to the local curvature and density that can avoid holes generation during the simplification process. Finally, the simplified points are triangulated by Cocone algorithm for surface reconstruction. This algorithm has been applied to a set of large scanned models. Experimental results demonstrate that it can generate high-quality surface approximation with feature preservation.