Point Clouds Segmentation for Complex Architectural Heritage

The poster was presented at the conference Artificial Intelligence for Cultural Heritage in 2019 and shows the cutting edge technologies for point clouds segmentatio and classifiction in the field of cultural heritage documentation. Different approaches were tested in application with standard scenarios:

  • the well known RANSAC algorithm for geometric primitives segmentation;
  • the CANUPO algorithm implementing a statistical approach for objects filtering;
  • the DoN algorithm implementing a normal based approach for filtering
Conference poster