A Study on the Automated Identification of Rock Fractures Using Yolov8

DOI

10.2991/978-94-6463-900-1_51How to use a DOI?

Keywords

Deep Learning; Rock Fracture; Automated Detection; YOLOv8

Abstract

The automatic detection of rock fractures is critical for maintaining the stability of rock structures and preventing geological disasters. Currently, three primary techniques are employed: manual inspection, ultrasonic detection, and computer vision technology. Manual inspection is inefficient, subjective, and poses significant safety risks in hazardous environments. Ultrasonic detection, while capable of providing accurate fracture data, has limited coverage and is therefore unsuitable for large-scale inspections. Computer vision techniques, which typically involve image preprocessing and edge detection to extract fracture contours, are often hindered by background noise and variations in lighting conditions.

To improve the efficiency and accuracy of rock fracture detection, this study proposes a high-performance automated detection algorithm based on the YOLOv8m-seg model. A dataset comprising 1051 annotated rock fracture images was collected and divided into training, testing, and validation sets in an 8:1:1 ratio. The YOLOv8m-seg model achieved an average precision (mAP@0.5) of 94.8%, demonstrating exceptional performance. Additionally, the algorithm exhibited high operational efficiency during deployment, with preprocessing requiring 2.5 milliseconds per image, inference 6.3 milliseconds, and post-processing 69.4 milliseconds. These results provide valuable insights for advancing research and practical applications in rock fracture detection.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Shao-Bin Xie
AU  - Fu-Hsuan Yeh
AU  - Shih-Heng Tung
PY  - 2025
DA  - 2025/12/07
TI  - A Study on the Automated Identification of Rock Fractures Using Yolov8
BT  - Proceedings of the Rocscience International Conference 2025 (RIC 2025)
PB  - Atlantis Press
SP  - 513
EP  - 520
SN  - 2589-4943
UR  - https://doi.org/10.2991/978-94-6463-900-1_51
DO  - 10.2991/978-94-6463-900-1_51
ID  - Xie2025
ER  -