Geological outcrop modelling and interpretation using ground based hyperspectral and laser scanning data fusion

Recent developments in the utilisation of close range laser scanning (lidar) in geology have seen the increased use of virtual outcrop data. The remote mapping of rock properties within the virtual outcrop remains, however, a challenge. This study aims to develop methods for combining and utilising data from close range lidar and ground based hyperspectral scanning. The workflow for using such data is presented for a case study of a dolomite and limestone quarry in the Peak District, UK. Multiple hyperspectral and lidar scans were acquired simultaneously to gain both spectral and geometric data. Different image processing algorithms were utilised to extract and map geological features from the spectral images, resulting in thematic images. To register the spectral images to the lidar coordinate system, the scanning sensor was modelled using cylindrical imaging geometry, as the central perspective projection is unsuitable for a rotating array. Bundle adjustment permitted the hyperspectral image orientation and position to be calculated, allowing the integration to be completed. The classified and registered imagery was used to colour the lidar point cloud, and this was superimposed with a textured 3D model of the outcrop, allowing interpretation and validation. The preliminary results showed areas of limestone and low dolomitization in the upper part of the quarry thought to have been only dolomite. This combination of geometrically accurate lidar data and spectral mapping of lithology has significant implications for the improved collection of geological data.