Automated house internal geometric quality inspection using laser scanning

Taking a terrestrial laser scanner to scan the room of a house, the scanned data can be used to inspect the geometric quality of the room. Taking advantage of the scan line feature, we can quickly calculate normal of the scanned points. Afterwards, we develop a fast plane segmentation approach to recognize the walls of the room according to the semantic constraints of a common room. With geometric and semantic constraints, we can exclude points that don’t belong to the inspecting room. With the segmented results, we can accurately do global search of max and min height, width and length of a room, and the flatness of the wall as well. Experiment shows the robustness of this geometric inspecting approach. This approach has the ability to measure some important indicators that cannot be done by manual work.