On the use of drones for vision-based digitalization, diagnostics and energy efficiency assessment of buildings

Abstract

The adoption of Unmanned Aerial Vehicles for energy efficiency assessment is a promising technique. Although there are some interesting ideas to automate the process, most of the operations are still done manually. An integrated methodology is presented to identify maintenance needs from both RGB and infrared images collected with a commercial drone. The real building selected as case study is reconstructed in a 3D environment through Structure-from-Motion Photogrammetry, while the infrared information is integrated and properly scaled on it to obtain a 3D point cloud model that provides, for each point: (i) information about its position in the space, and (ii) external surface temperature measured during data gathering phase. The point cloud is then segmented into the different sides of the envelope to identify each part of the façade and compared with a model that reconstructs the expected conditions of the building. This procedure enables the development of an automated pipeline to identify defects and failures in a building envelope and to suggest corrective actions