Autonomous scanning of structural elements in buildings

Abstract

This paper presents a method for automatic scanning of structural elements in indoors. Whereas most of the next-best-scan (NBS) based methods do not separate clutter and useful data, we present a scanning strategy in which potential structural components (SE) of the building are recognized as a new scan is carried out. This makes our method more efficient and less time consuming compared with the rest. Besides, our approach gives a response to essential issues in the scanning world, such as the data discrimination, the hypotheses about the workspace and the complexity of the scanned scene. The method has been tested in indoors under occlusion and clutter yielding promising results. Additionally, a comparison with three techniques close to ours is included in the paper.