Technology and Methodology for Obtaining 3D Omnidirectional Models of Thermal Point Clouds of Buildings

Abstract

To date, there is scarce research on the creation of 3D Semantic Thermal Models of buildings (3D-STM) using long-range sensors. This research line opens up the field of applications in the world of building energy performance, the main issue in current European research guidelines. This paper presents a new technological platform and methodology that generates an omnidirectional thermal point cloud, as a previous stage of the creation of 3D-STM. The system integrates a large-range laser scanner, a thermal camera, and a motorized pan-tilt head. The system is programmed to capture omnidirectional 3D thermal data of the scene, with wide horizontal and vertical ranges. After carrying out several thermal scans of the scene (a room or a story), a whole 3D omnidirectional thermal point cloud is automatically obtained. In this process, several problems are presented and solved, among others, those derived from the variability of the thermal camera, topics which are scarcely treated in the existing bibliography. This work considerably reduces the thermal noise introduced. The final aim of this project is to integrate 3D-STM with BIM, having a direct impact on the development of future intelligent buildings and on the new energy efficiency paradigms.