Développement d'une méthode multi-échelle de traitement des nuages de points LiDAR mobile pour la détection de convergence des tunnels miniers souterrains
|Advisor:||Mostafavi, Mir Abolfazl; Larouche, Christian|
|Abstract:||Security in underground mines is one of the main concerns of mining companies. Monitoring any changes or distortions in mining tunnels and galleries, including tunnel’s convergence, is one of the top priorities for those companies. Commonly used convergence monitoring techniques in underground tunnels are based on limited measurements that are usually very time-consuming and costly in terms of process. Emerging mobile LiDAR technologies offer alternative solutions that are far more efficient and lower in risk. However, the accurate measurements of convergence in mining tunnels using mobile LiDAR data remains challenging because of the following issues: limited precision of LiDAR point clouds, the roughness and irregularity of the studied surfaces in the tunnels, the irregularity of the scan density, the presence of occlusions and the complexities of reliable registration of point clouds. This research work is defined as a part of the ‘MinEyes project’ in collaboration with the company Pecktech which main objective is to detect and measure the convergence of underground mining tunnels from mobile LiDAR data. To do so, we propose a multi-scale point cloud comparison method that considers the surface irregularities of the mining tunnels and allows a better detection of their changes and a more precise measurement of their convergence. The proposed method is based on a local model-to-model comparison method of underground mobile LiDAR point clouds. The results obtained are very promising compared to other LiDAR point cloud comparison methods (e.g. C2C, C2M methods). A statistical analysis of the results also has confirmed the efficiency of the proposed method, especially in the rough areas of the mining tunnels.|
|Document Type:||Mémoire de maîtrise|
|Open Access Date:||13 March 2019|
|Collection:||Thèses et mémoires|
All documents in CorpusUL are protected by Copyright Act of Canada.