Abstract Jq1 n0O
The collection of geomechanical data in the mining industry is often hindered by limited access to rock exposures. Consequently, "kSwa16O
there is a need for a tool that permits fast and safe acquisition of information that best characterizes a geological structural regime. d<T%`:s<
This paper presents a digital face mapping methodology used to construct discontinuity trace maps from photographs of rock faces. \ lr/;- zP
The method is applicable under a range of ground conditions, while at the same time trying to keep user intervention to a minimum. (cV
The developed procedure uses a series of photographs of a rock face taken under symmetrical lighting orientations. All digitized bx;f`8SN
pictures are then introduced into the developed software, where pre-processing optimizes subsequent analysis. Discontinuity traces &=w|vB)(p
are then extracted from the enhanced images using edge and line detection algorithms. This results in a binary image where black UzQ$B> f
pixels correspond to the detected features. The nature of these features is identified using artificial neural networks, specifically ;h=S7M9.
trained for this application. As a result, it is possible to distinguish between segments in the image that are due to the presence of a (_8#YyW#
discontinuity and those due to other phenomena. This classification process provides a series of binary images that represent sBjXE>_#)
discontinuity traces in the area of interest. These images are superposed and merged. All processes described above are fully :YvbU Y
automatic. In order to complete the process, the system operator ensures that incomplete discontinuity segments are reconstructed. I,P!@
Following this, a discontinuity trace map is constructed and the discontinuity network is characterized using geomechanical criteria. &YX6"S_B
The methodology to construct discontinuity trace maps is illustrated using an example from a rock exposure in an underground VXC4%
mine. Subsequently the methodology is implemented in two case studies from underground mines and one in a quarry. These case )/Y~6A9>
studies represent different geological and lighting conditions. This information is then used to successfully characterize the .h8M
investigated rock masses. Present limitations of the method are also identified. CT"Fk'B'
r 2003 Elsevier Ltd. All rights reserved.
