Instance segmentation for geotechnical core analysis in mining

Thumbnail Image

Files

master_Tan_Lucas_2025.pdf (6.08 MB)

URL

Journal Title

Journal ISSN

Volume Title

School of Engineering | Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.

Authors

Date

2025-11-24

Department

Major/Subject

Mining Engineering

Mcode

Degree programme

European Mining Course

Language

en

Pages

92

Series

Abstract

Geotechnical core logging lies at the foundation of mine design, yet it remains laborious, subjective, and prone to methodological lock-in. Meanwhile, core box images represent a prevalent and largely untapped primary data source. Recent studies have shown that these images can be used to derive RQD by training an instance segmentation algorithm, a deep learning method that can distinguish between each individual core piece. This thesis presents the first zero-shot approach to instance segmentation by using Meta’s Segment Anything Model (SAM), which requires no retraining and offers the potential for rapid deployment across sites. It evaluates the technical performance and the practical realities of integrating this novel approach when applied to operational datasets. This demo was created as a QA/QC tool to verify manual RQD logs. The methodology applies image preprocessing, SAM inference, and post-processing to a pre-existing picture database at a major European base metals mine. Using this mine as a case study allows us to contextualize this methodology by considering the operational constraints of an active mining environment. It also allows assessment of the data quality in legacy datasets. The methodology accurately segmented intact core pieces, achieving a MAE of 7.7 percentage points, demonstrating alignment between automated and human-derived results. Outliers demonstrated shortcomings in the input data and errors during the manual logging process, rather than errors with the application. Actual errors appeared due to failure to detect core pieces (false negatives), failure to separate multiple core pieces (undersegmentation) and the inability to filter out certain artefacts (false positives). An important limitation is the inability to distinguish natural from artificial fractures. Improving the consistency of core box photographs would improve OCR and segmentation accuracy. This work demonstrates that SAM-based zero-shot segmentation can form a practical foundation for extracting fracture data from core box photographs, when combined with filtering steps based on individual and relational mask geometry and targeted removal of unwanted masks.

Description

Supervisor

Rinne, Mikael

Thesis advisor

Janiszewski, Mateusz

Keywords

Segment Anything Model, instance segmentation, rock quality designation, RQD, core box images, geotechnical logging, machine learning

Other note

Citation

Permanent link to this item

https://urn.fi/URN:NBN:fi:aalto-202512159188

Collections

[dipl] Insinööritieteiden korkeakoulu / ENG