- Title
- Rock mass trace line identification incorporated with grouping algorithm at tunnel faces
- Creator
- Leng, Biao; Yang, Hui; Hou, Gaopeng; Lyamin, Andrei
- Relation
- Tunnelling and Underground Space Technology Vol. 110, Issue April 2021, no. 103810
- Publisher Link
- http://dx.doi.org/10.1016/j.tust.2021.103810
- Publisher
- Elsevier
- Resource Type
- journal article
- Date
- 2021
- Description
- Geological trace lines including rock joints and fractures are important characteristics of tunnel faces. Fully recording and analyzing this information is very helpful for evaluating the state of the tunnel from engineering geology prospective. Based on the digital images of a tunnel face, the paper studied rock mass trace line extraction and grouping algorithms with the aim of minimizing the degree of manual intervention. Firstly, the characteristics of rock mass edges are analyzed, and then Canny operator is proposed for appraising trace lines of different types of rock masses. The algorithm applies image thinning, edge fitting, splitting, merging and filtering to all detected trace lines. On the basis of the statistical characteristics of the grey values of the pixels on both sides of the trace line in the original image, a one-sided expansion method is developed to dilate the trace line. After filling the expanded connected area, image thinning, polyline fitting, splitting, merging and filtering are performed again to obtain the final trace line of the rock mass exposure. Then, tunnel face characteristic extraction algorithms are presented to calculate rock mass characteristics, and a special clustering technique is developed to automatically group the trace lines. Applying this approach to rock mass images of the tunnel face may eliminate the double boundary lines detected on some edges and it also allows to perform the automatic characteristic extraction and grouping of trace lines, and consequently increase analysis automation of the tunnel face. The experimental results illustrate that the proposed method is effective for analysis and evaluation of the rock mass on the tunnel face.
- Subject
- tunnel face; image processing; rock mass; trace line
- Identifier
- http://hdl.handle.net/1959.13/1435153
- Identifier
- uon:39626
- Identifier
- ISSN:0886-7798
- Language
- eng
- Reviewed
- Hits: 1073
- Visitors: 1073
- Downloads: 0
| Thumbnail | File | Description | Size | Format |
|---|