Architecture Engineers Award

High-fidelity textured 3D models have become essential for documenting structural damage, guiding restoration decisions, and enhancing visual communication in architectural heritage conservation. Traditional LiDAR-based texture modeling is often impractical for large-scale sites due to their vastness, terrain complexity, and gaps in structural capture. Recent research explores alternative reconstruction approaches that overcome these limitations while maintaining high geometric accuracy and visual quality. Limitations of LiDAR-Based Texture Modeling in Large-Scale Heritage Sites Large architectural heritage sites often suffer from incomplete LiDAR scans caused by occlusions, irregular topography, and scale constraints. These limitations reduce the reliability of LiDAR-derived textures and restrict their role in conservation workflows. Understanding these challenges highlights the need for alternative acquisition and modeling techniques that can capture complex geometries more efficie...

Building Information Modeling (BIM) has revolutionized the architecture, engineering, and construction (AEC) industry, but the process of creating BIM models from unstructured point cloud data remains highly manual and time-consuming. Traditional methods rely on complex segmentation and calibration techniques, often requiring significant user intervention. Cloud2BIM introduces a breakthrough by automating this transformation process through advanced algorithms that segment walls, slabs, openings, and rooms, delivering a fully automated and accurate workflow for BIM model generation. Background and Motivation The rapid growth of 3D laser scanning and photogrammetry technologies has enabled the collection of detailed geometric data of buildings. However, converting this raw, unstructured point cloud data into usable BIM models continues to be a major challenge due to irregular geometries, noise, and the need for manual processing. The motivation behind Cloud2BIM lies in overcoming thes...