Automated 2D Measuring of Interiors using a Mobile Platform

This paper presents an approach for a fully automated measuring of a self-contained indoor scene in 2D using a mobile platform. A novel sensor is used, designed for the acquisition of precise floor plan scans, mounted on an ordinary mobile robot. The software framework is configured to optimize the accuracy and the completeness of the acquired data. Therefore an exploration strategy is presented that finds gaps in the data and determines the next best view in an explorative way. The scene is presented to the system as continuous line segments. Data management builds up the gathered point cloud in a homogeneous density, using the relevant and most accurate available information. Existing scan matching techniques are modified in a way to work robust and precise in the scenarios expected. The framework is tested in an exemplary indoor scene. Additionally, the scene is surveyed with a referencing system to build ground truth data and to enable an accuracy evaluation of the developed system.