Search for: All records

Building facade inspections are crucial for ensuring structural integrity and occupant safety, traditionally requiring physical access that can be costly and risky. Recent advancements have seen the integration of Unmanned Aerial Systems (UASs) equipped with imaging technologies to facilitate these inspections. However, the primary challenge remains in the effective detection and precise localization of facade anomalies, such as cracks, stains, and specifically, thermal anomalies. This paper investigates the fusing of diverse data sources, specifically thermal infrared (IR) imaging and high-resolution RGB data, to enhance 3D registration of thermal anomalies. Utilizing a Computational Neural Radiance Field (NeRF) approach, it aims to reconstruct detailed 3D models of building facades. This integration improves the accuracy of anomaly detection by fusing the precise camera positions derived from the original RGB data to refine the alignment and visualization of thermal anamolies in both RGB and IR imagery. We utilized COLMAP for camera position estimation and NeRF for 3D registration, employing Structure-from-Motion (SfM) and NeRF methodologies to create detailed and scalable 3D models from 2D IR and RGB images. This approach integrates the precision of camera positioning with advanced 3D reconstruction techniques to enhance the visualization and analysis of building facades. Our method automates the registration and mapping of detected thermal anomalies in 2D images onto the reconstructed 3D models, improving the diagnostics of building facades by enabling precise localization and scaling of these anomalies. The findings demonstrate the potential of our approach to reduce inspection times and enhance the safety of diagnostic procedures through higher accuracy and less invasive methods. 

Buildings account for 40% of energy consumption in the U.S., with significant energy losses stemming from poor insulation and leaks. To meet decarbonization goals by 2050, efficient methods for diagnosing and retrofitting thermal anomalies are essential. Infrared thermography (IRT), combined with emerging technologies like computer vision and deep learning, offers the potential for automating thermal anomaly detection and segmentation. However, challenges such as building diversity, scenario variations, and labor-intensive image annotation hinder model reliability and robustness. This study proposes a human-in-the-loop active learning approach to fine-tune the pretrained Capsule-based network (CapsLab) to enhance adaptability to new building scenarios efficiently. Using a Query-by-Committee (QBC) strategy, the method selects the most informative thermal images from a target dataset for expert verification and iterative model refinement. To streamline the expert annotation verification process, a weakly annotation strategy is introduced supporting human-in-the-loop training by Simple Linear Iterative Clustering (SLIC)-based superpixel segmentation and scribble prompts for efficient labeling. The pretrained CapsLab model is iteratively fine-tuned using these verified annotations to enhance segmentation performance. Three fine-tuning methods, including parameter freezing strategies, are evaluated for optimal results. This iterative workflow reduces annotation effort, improves model adaptability, and improves the precision of thermal anomaly segmentation, facilitating energy-efficient building retrofitting and maintenance.