Using a linear octree to identify empty space in indoor point clouds for 3D pathfinding

Indoor pathfinding and routing need fast ways to define connected navigable space, which represents usable paths. This is important because the interior space in buildings changes and often does not follow the architectural design. A workflow is presented that uses a linear octree to segment the space contained in an indoor point cloud, and subsequently derives the structured and connected empty space. Binary locational codes are efficiently generated for every point, which implicitly contain the linear octree structure. The workflow successfully generated the linear octree and derived the empty space. A shortest path algorithm showed the potential of using a linear octree for space segmentation in indoor point clouds, and deriving the structured and connected empty space. The workflow is fast and automated, enriches point clouds with a structure that can be understood by computers and retains all detail of the raw data set.