Identification and Visualization of Unscanned Areas Within a Building Based on the Building’s Outer Hull

Abstract. 3D evaluation of indoor environments is important for first responders and their commanders. Both are unfamiliar with the situation that they will find. While doing their work they can scan the environment they are working in and process that data in real time. To be able to do this efficiently both the scanned areas and the unscanned parts need to be assessed giving indications where to go to next. In this research an approach to identify, classify and visualize unscanned areas within the building is proposed. Using RGB and Depth (RGB-D) cameras and Collaborative Visual SLAM a pointcloud of the interior of the building is generated. By combining the pointcloud with the building’s outer boundary it is possible to evaluate which areas have been visited by the responders, which have not and identify potential hidden spaces. The proposed solution implements that by employing voxels. The result can be seen on the online 3D web-viewer. The research shows promising results but also aspects that will have to be improved like the alignment of indoor and outdoor parts. Also the narrow field of view of the used Li- DAR device (iPhone) brought limitations.