Visual Enhancement of 3D Images of Rock Faces for Rock Mass Characterization

This paper investigates how 3 dimensional (3D) laser images of rock faces can be processed to highlight several characteristics of the rock mass in three different geological environments. The first scene features large cubic quartzite rock blocks from a road cut. The second scene shows a layered sedimentary sequence composed of massive limestone and more friable dolostone from a quarry. The third scene exhibits shatter cone structures imprinted in an outcrop in the Subdury impact crater, in northern Ontario. Images are acquired using a triangulation laser camera with a 30° × 30° field of view, mounted on a tripod, at distances ranging from 1 to 3 m. Epsilon nets (e-nets) are used to downsample the point cloud data with little loss to the overall structure that they represent. A 2.5D triangular irregular mesh (TIN) is constructed from the 3D point cloud data using a delaunay triangulation and smoothed using a few iterations of the geometric laplacian. Three different processing strategies are tested to enhance the visual perception of fracture orientation and surface roughness: color-coding of strike and dip, curvature analysis and outer normal planar clustering.