Abstract
The demand for three-dimensional (3D) urban model is growing in Japanese traditional buildings for disaster prevention. Classification of point clouds into structural elements is required for generating 3D models. This paper proposes an algorithm to automatically extract the lattice windows of Japanese traditional buildings from terrestrial LiDAR data for the sake of 3D modeling. First, it extracts edges from three dimensional data. Second, the edges are classified into groups by distance between two points. Finally, lattice windows are extracted from the groups using SVM (support vector machine) classifier. The validation results showed that the accuracy to finally extract lattice windows was 63%. As future tasks, edge extraction processing should be improved, and better features for SVM classifier should be developed.
