SfM Meets Thermal: Reconstructing 3D Thermal Profiles of Buildings Using Multimodal Drone Data

Heat loss in buildings is a growing concern in the US, driven by rising electricity costs and an aging housing stock, with 44% built before 1970. Poor insulation in older homes increases energy consumption, costs, and environmental pollution. Thermal renovation, guided by accurate 3D thermal profiles, is vital for effective energy inspection and decision-making. Drone imagery offers an accessible solution, but standard drones lack high-resolution thermal sensors for detailed profiling. This work presents a novel method to fuse low-resolution thermal data with high-resolution visible data, resulting in a high-resolution 3D thermal profile of buildings using open-source structure-from-motion (SfM) pipelines. Testing on an academic building at the University of Texas at Arlington demonstrated significant resolution improvements, enabling precise identification of thermal anomalies. This advancement enhances energy efficiency assessments and identifies heat loss zones more effectively. Future work will refine this method for broader and more robust applications.