660 likes | 677 Views
Alexander Hornung and Leif Kobbelt RWTH Aachen. Robust Reconstruction of Watertight 3D Models from Non-uniformly Sampled Point Clouds Without Normal Information. Point Cloud Reconstruction. Point Cloud Reconstruction. Non-uniform sampling Holes Noise Bad scan alignment
E N D
Alexander Hornung and Leif Kobbelt RWTH Aachen • Robust Reconstruction of • Watertight 3D Models from • Non-uniformly Sampled Point Clouds Without Normal Information
Point Cloud Reconstruction • Non-uniform sampling • Holes • Noise • Bad scan alignment • No (reliable) normals
Point Cloud Reconstruction • Smooth watertight manifold • No topological artifacts (low genus) • Detail preservation • Robustness to • Non-uniform sampling • Holes • Bad registration and noise • From 3D points only
Outline • Introduction • Surface confidence estimation • Graph-based surface extraction • Hole filling and detail preservation • Mesh extraction • Results
Related Work • Wrapping and Voronoi-based • Amenta et al., Bernardini et al., Boissonat and Cazals, Dey and Goswami, Mederos et al., Scheidegger et al., … • Deformable models • Esteve et al., Sharf et al., … • Volumetric reconstruction • Hoppe et al., Curless and Levoy, Carr et al., Ohtake et al., Fleishman et al., Kazhdan, …
Related Work • Wrapping and Voronoi-based • Amenta et al., Bernardini et al., Boissonat and Cazals, Dey and Goswami, Mederos et al., Scheidegger et al., … • Deformable models • Esteve et al., Sharf et al., … • Volumetric reconstruction • Hoppe et al., Curless and Levoy, Carr et al., Ohtake et al., Fleishman et al., Kazhdan, … • Graph-based energy minimization and surface reconstruction • Boykov and Kolmogorov, Vogiatzis et al., Hornung and Kobbelt
Overview • Point cloud P
Overview • Point cloud P • Surface confidence (unsigned distance)
Overview • Point cloud P • Surface confidence (unsigned distance) • Embed weighted graph structure G
Overview • Point cloud P • Surface confidence (unsigned distance) • Embed weighted graph structure • Min-Cut of G yields unknown surface
Outline • Introduction • Surface confidence estimation • Graph-based surface extraction • Hole filling and detail preservation • Mesh extraction • Results
Surface Confidence • Insert 3D samples into volumetric grid • Sparse set of occupied voxels • Compute a confidence map “Probability” that surface intersects a voxel v
Surface Confidence • Insert 3D samples into volumetric grid • Sparse set of occupied voxels • Compute a confidence map “Probability” that surface intersects a voxel v • Compute “crust” containing the surface • Morphological dilation • Medial axis approximation
Surface Confidence • Insert 3D samples into volumetric grid • Sparse set of occupied voxels • Compute a confidence map “Probability” that surface intersects a voxel v • Compute “crust” containing the surface • Morphological dilation • Medial axis approximation • Estimate by volumetric diffusion
Outline • Introduction • Surface confidence estimation • Graph-based surface extraction • Hole filling and detail preservation • Mesh extraction • Results
Find Optimal Surface • Minimize energy • Min-Cut of an embedded graph • Global optimum • Highly efficient • Graph structure?
Dual Graph Embedding • : Probability that v is intersected by surface s • Intersected voxels are split into 2 components • Interior faces • Exterior faces
Dual Graph Embedding Voxel split-edges Graph cut-edges • : Probability that v is intersected by surface s • Intersected voxels are split into 2 components • Interior faces • Exterior faces Split along a sequence of edges • Octahedral graph structure
Min-Cut Surface Extraction • Embed graph into a crust containing the surface
Min-Cut Surface Extraction • Embed graph into a crust containing the surface • Edge weights defined per voxel
Min-Cut Surface Extraction • Embed graph into a crust containing the surface • Edge weights defined per voxel • Min-cut yields set of intersected surface voxels
Min-Cut Surface Extraction • Embed graph into a crust containing the surface • Edge weights defined per voxel • Min-cut yields set of intersected surface voxels • Parameter s to emphasize strong/weak maxima
Outline • Introduction • Surface confidence estimation • Graph-based surface extraction • Hole filling and detail preservation • Mesh extraction • Results
Hierarchical Approach • Single resolution impractical • High volumetric resolutions • Non-uniform sampling / large holes • Hierarchical framework • Adaptive volumetric grid (Octree) • Proper initial crust at low resolutions • Simple narrow-band approach insufficient • Loss of fine details not contained within crust
Hierarchical Approach • Single resolution impractical • High volumetric resolutions • Non-uniform sampling / large holes • Hierarchical framework • Adaptive volumetric grid (Octree) • Proper initial crust at low resolutions • Simple narrow-band approach insufficient • Loss of fine details not contained within crust Re-insertion of data samples • Merge samples with crust
Hierarchical Approach • Surface confidence estimation • (Re-)Insert point samples • Dilate and compute • Graph-based surface extraction • Generate octahedral graph • Compute min-cut • Volumetric refinement • Narrow band 643
Hierarchical Approach • Surface confidence estimation • (Re-)Insert point samples • Dilate and compute • Graph-based surface extraction • Generate octahedral graph • Compute min-cut • Volumetric refinement • Narrow band 1283
Hierarchical Approach • Surface confidence estimation • (Re-)Insert point samples • Dilate and compute • Graph-based surface extraction • Generate octahedral graph • Compute min-cut • Volumetric refinement • Narrow band 1283
Hierarchical Approach • Surface confidence estimation • (Re-)Insert point samples • Dilate and compute • Graph-based surface extraction • Generate octahedral graph • Compute min-cut • Volumetric refinement • Narrow band 2563
Hierarchical Approach • Surface confidence estimation • (Re-)Insert point samples • Dilate and compute • Graph-based surface extraction • Generate octahedral graph • Compute min-cut • Volumetric refinement • Narrow band 2563
Hierarchical Approach • Surface confidence estimation • (Re-)Insert point samples • Dilate and compute • Graph-based surface extraction • Generate octahedral graph • Compute min-cut • Volumetric refinement • Narrow band 5123
Hierarchical Approach • Surface confidence estimation • (Re-)Insert point samples • Dilate and compute • Graph-based surface extraction • Generate octahedral graph • Compute min-cut • Volumetric refinement • Narrow band 5123
Outline • Introduction • Surface confidence estimation • Graph-based surface extraction • Hole filling and detail preservation • Mesh extraction • Results
Cut Manifold to Triangle Mesh Graph cut-edges Loop of voxel split-edges
Cut Manifold to Triangle Mesh • Loops define non-planar polygonal faces • Mesh vertices at voxel corners
Cut Manifold to Triangle Mesh • Loops define non-planar polygonal faces • Mesh vertices at voxel corners • Cycle along split-edges
Cut Manifold to Triangle Mesh • Loops define non-planar polygonal faces • Mesh vertices at voxel corners • Cycle along split-edges
Cut Manifold to Triangle Mesh • Loops define non-planar polygonal faces • Mesh vertices at voxel corners • Cycle along split-edges
Cut Manifold to Triangle Mesh • Loops define non-planar polygonal faces • Mesh vertices at voxel corners • Cycle along split-edges
Cut Manifold to Triangle Mesh • Loops define non-planar polygonal faces • Mesh vertices at voxel corners • Cycle along split-edges
Cut Manifold to Triangle Mesh • estimated per voxel Mesh vertices at voxel centers
Cut Manifold to Triangle Mesh • estimated per voxel Mesh vertices at voxel centers
Cut Manifold to Triangle Mesh • estimated per voxel Mesh vertices at voxel centers • Voxel corners correspond to non-planar faces • Cycle over shared split-edges
Cut Manifold to Triangle Mesh • estimated per voxel Mesh vertices at voxel centers • Voxel corners correspond to non-planar faces • Cycle over shared split-edges
Cut Manifold to Triangle Mesh • estimated per voxel Mesh vertices at voxel centers • Voxel corners correspond to non-planar faces • Cycle over shared split-edges