AAPG Annual Convention and Exhibition

Datapages, Inc.Print this page

3-D Outcrop Modelling Using UAV-Based Photogrammetry: Comparison to Terrestrial Lidar Methods

Abstract

Digital outcrop modelling using laser scanning (lidar) has become increasingly widespread over the last decade, with the high resolution and completeness allowing the geometry of features to be constrained with high accuracy. The result of laser scanning is a 3D point cloud with additional digital images that are used to create photorealistic 3D outcrop models (virtual outcrops), which can then be analysed quantitatively and qualitatively. Photogrammetry is a complementary method that uses overlapping imagery alone to determine 3D object positions. Central to the technique is the identification of common features between sets of overlapping images, which are then used as the basis for reconstructing each image's orientation and position at the time of exposure, simultaneously with the 3D position of the feature points and camera calibration information. This stage results in a sparse point cloud and the orientation parameters of all used images. Subsequently, the sparse point cloud is densified by performing point matching at the pixel level, potentially resulting in point clouds with similar densities as lidar data. The final result is largely identical to application of a lidar system: dense point clouds and associated imagery which can be used to create virtual outcrop models. In this research, an unmanned aerial vehicle (UAV) was used as the camera platform for obtaining image sets for photogrammetric modelling. Data were collected for two areas in northern Spain: the Ainsa 1 turbidite system exposed in the Ainsa quarry, and fluvial deposits from the Huesca Fan in outcrops below the town of Bolea. Both outcrops were difficult to access for conventional fieldwork. Imagery were collected and processed to form dense 3D point clouds, which were then meshed and textured to form photorealistic models. Georeferencing was carried out using an on-board GPS, which was used to tag the individual images. A comparison with terrestrial lidar was performed to highlight the success and independence of the method. Results suggest that photogrammetry is a valuable alternative and low-cost technique for obtaining 3D outcrop data in many situations, and the UAV platform offers a number of benefits over ground-based data capture. The advantages and limitations with respect to conventional lidar methods are discussed. Finally, the virtual outcrops were used to build reservoir models for investigating facies architecture.