Room Map Estimation from Two-Dimensional Lidar's Point Cloud Data

Abstract

An automatic digital map estimation is not only essential for the repository of the architectural plans but also crucial for providing clear paths for a trajectory finder system, an autonomous vehicle, or a robot. Therefore, it is useful for reconstructing heritage buildings or finding a location lying in ruin after an earthquake. Rebuilding a digital map of a room can be done by using information from two-dimensional (2D) images. However, it seems to be complicated to obtain an accurate dimension directly from those images. One potential breakthrough is to make use of light detecting and ranging (LIDAR) technology. This state-of-the-art device can detect the solid surface of an object, as well as presenting an easiness of the distance measurements. This research focused on the reconstruction of a room map by utilizing 2D point cloud data obtained from a Lidar device. The results showed that by applying the conditioned random sample consensus (RANSAC) method, the 2D map of a room could be identified accurately from the 2D point cloud data.