@phdthesis{Fotsing2024, author = {Fotsing, Cedrique}, title = {Point cloud processing : 3D indoor scans of buildings as essential tools in engineering, design and construction}, doi = {10.26127/BTUOpen-6894}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus4-68947}, school = {BTU Cottbus - Senftenberg}, year = {2024}, abstract = {Recent advances in laser scanning systems have enabled the acquisition of 3D point cloud representations of scenes. This technological progress has significantly transformed and revitalized several sectors, including architecture, engineering, and construction (AEC) fields. The fundamental purpose of this thesis is to design and propose a set of methods for the automatic generation of 3D semantic models of multi-level buildings from indoor point clouds. The research encompasses a comprehensive workflow, starting with the acquisition of point cloud data, progressing through essential segmentation techniques for identifying distinct architectural elements, and culminating in the generation of the final model. The initial part of our work provides a comprehensive overview of diverse approaches to 3D scanning and technologies. Subsequently, we introduce an innovative point cloud registration technique, essential in obtaining both high-quality and comprehensive point cloud datasets. In the second part, we mainly focus on segmentation processes. A method for extracting planar structures is developed, followed by a specialized method for segmenting wall structures and identifying openings in wall blocks. The third part mainly focuses on the presentation of BIM technologies and their usefulness in the fields of AEC. Finally, the last part is devoted to the design of a complete pipeline for generating the 3D model from indoor point clouds. In the proposed pipeline, after segmenting the point clouds into potential building floors, a wall detection process is performed on each floor segment. Then, room, ground, and ceiling extractions are conducted using the 2D constellation obtained from the projection of the walls onto the ground plan. The identification of the openings in the walls is performed using a deep learning-based classifier that separates doors and windows from non-consistent holes. Based on the geometric and semantic information from previously detected elements, the final model is generated in IFC format. The effectiveness and reliability of the proposed methods and pipeline are demonstrated through extensive experiments at each step. The results reveal high precision and recall values, ensuring the fidelity of the generated models. The pipeline's efficiency and accuracy provide valuable contributions to future advancements in point cloud processing.}, subject = {Building Information Modeling; 3D scanning; Indoor point cloud; 3D-Scanning; Indoor-Punktwolke; Point cloud segmentation; 3D semantic models; Punktwolkensegmentierung; 3D-semantische Modelle; 3D-Scanner; Dreidimensionales Modell; Geschossbau; Punktwolke; Segmentierung; Building Information Modeling}, language = {en} }