Potentials of RGB-D Cameras in As-Built Indoor Environment Modeling

3D as-built models of building indoor environments could be used to facilitate multiple building assessment and management tasks, such as post-disaster safety evaluation, renovation/retrofit planning, and maintenance scheduling. However, modeling building indoor environments is a challenging task. It is even more difficult than modeling building facades due to the issues, such as limited lighting conditions and prevalence of texture-poor walls, floors, and ceilings. This paper investigates the potentials of RGB-D cameras in modeling building indoor environments. Three pilot studies have been performed to evaluate 1) the accuracy of the sensing data provided by an RGB-D camera and 2) the automation that can be achieved for the registration of building indoor scenes and the recognition of building elements with the sensing data. The studies show that the camera can provide a stream of mid-accurate sensing data in real time. Also, a high degree of automation can be achieved through the fusion of spatial and visual data from the camera, when modeling the as-built conditions in building indoor environments.