Implicit Surface Reconstruction Based on Adaptive Clustering

This paper presents an efficient method to reconstruct an implicit surface from a point cloud based on adaptive clustering. First, a novel approach, based on a hierarchical clustering tree with variation and normal constraints, is proposed for normal estimation of the point cloud. It simplifies the point cloud and reduces the amount of direction propagation paths in normal orientation. Second, an implicit surface is computed from the clustering tree by a feature preserving method, using adaptive plane fitting that considers the local feature and sharp areas of the point cloud. Finally, the results demonstrate the efficiency of our method for open, closed and unorganized models, and show that the resulting surfaces preserve the sharp features.