Objectives According to the spatial attitude characteristics of the indoor walls, a point cloud vector tracking algorithm for automatically drawing the indoor plan was proposed.
Methods Firstly, the indoor wall point cloud with a certain thickness was intercepted and projected onto the horizontal plane, and square grids were established on the projection plane. The plane points after projection were segmented and managed, and the point cloud in every grid was clustered by the eight neighborhood algorithm. Then the grid center of gravity method was used to generate sparse points, which were tracked and sorted according to the distance between points in the neighborhood and the angle of connection vector. The vector sequence of tracking direction was established, and the point cloud segmentation of different walls was completed combined with mutation of angle between adjacent values in the vector sequence. Finally, the least squares algorithm was used to fit the line of the same wall point cloud, and the intersection point of the adjacent wall line was obtained.
Results The plane projection line segment of the house wall and its spatial connection relationship were established. The wall projection line segment was output in sequence. The drawing of the house plan was completed and the DXF format data exchange file was exported.
Conclusions The proposed algorithm was verified by the analysis of indoor scanning data of a residential building. The results show that the proposed algorithm can accurately and quickly analyzed and processed the indoor laser scanning point cloud, and completed the drawing of interior plan.
DownLoad: