Algorithm for automatic extraction of information from 3D data for underground mining tunnel models

Research output: Types of thesis › Master thesis

Contributors

  • Ahmad Elalailyi - , Polytechnic University of Milan (Author)

Abstract

The mining industry has been proliferating through the past decades, showing a great impact on economies, societies, and industries' development. In particular, underground mining is characterized by harsh environments and thus should be undergone with extreme caution to prevent any intolerable consequences. Therefore, imposing regular and repetitive inspections and monitoring to detect risks, ensure structural stability, and extend tunnel life. Visual inspection has been one of the most desirable and valuable non-destructive methods for providing preliminary information on the structure's condition. However, traditionally this required the physical presence of trained inspectors examining large spans with traditional surveying and monitoring instruments that are time-consuming and error-prone due to human subjectivity. At present, the modern surveying field especially Mobile Mapping Systems is growing and being adopted widely, but handling various forms of 3D data still encounters some challenges that constrain its potential benefits. Accordingly, mining development requires both advancement in surveying technology and the effective utilization of data to ensure efficient productivity and positive environmental performance. This paper proposes an algorithm taking into consideration the needs and requirements of one of the most prestigious mining companies in Italy based on an interview conducted with them to develop a fast, automatic, least time, and effort-consuming optimal workflow where 3D data can assist professionals in studying the tunnels environment. It's capable of (1) skeletonizing the tunnel and describing its geometry and topology by creating a simplified shape descriptor; (2) extracting cross-sections along the central axis; (3) referencing and extracting images at specified locations describing the general scene of the working environment; (4) performing tunnel decomposition and measure various geometric parameters. The results demonstrate that data processing can be automated in an innovative, efficient, and unchallenging manner, allowing mine owners to display, maintain, process, and analyze large amounts of 3D data with the least requirements, increasing their operations' safety, productivity, and sustainability.

Details

Original languageEnglish
Qualification levelMaster of Science
Awarding Institution
  • Polytechnic University of Milan
Supervisors/Advisors
  • Fassi, Francesco, Supervisor, External person
Defense Date (Date of certificate)28 Apr 2022
Publication statusPublished - 28 Apr 2022
Externally publishedYes
No renderer: customAssociatesEventsRenderPortal,dk.atira.pure.api.shared.model.researchoutput.Thesis

External IDs

ORCID /0009-0000-2012-1203/work/219268437

Keywords