Wire mesh filtering in 3D image data of rock faces

Abstract

The presence of wire mesh partially obscures rock faces imaged using terrestrial laser scanners. The objective of the methodology presented in this paper is to remove the wire mesh from point cloud data, leaving rock face data suitable for joint orientation and surface roughness analysis. Principal Component Analysis (PCA) is first performed to segregate data belonging to the wire mesh from legitimate rock face data. Two iterative filters are then applied. The baseline filter relies on an estimation of the instrument noise and surface roughness. Rockbolt face plates are used to measure the instrument noise in situ. A second derivative filter based on a central finite difference approximation is used to identify the boundary between rock face and wire mesh data. This filtering methodology was successful in removing wire mesh and associated edge effects from point cloud data allowing for joint orientation and surface roughness assessment, the latter of which was found to be more sensitive to the presence of the wire mesh.