New Approaches to Data Collection and Analysis Using Virtual-Reality Based Data Visualization Tools 3D Visualizer, Crusta, and Lidar Viewer at Keckcaves

Adam M. Forte^1,2, Eric Cowgill^1,2, Louise H. Kellogg^1,2, Mike Oskin^1,2, Oliver Kreylos^1,2, Tony Bernardin^3,4, Magali I. Billen^1,2, Peter Gold^1,5, Margarete Jadamec^1,6, Tyler Mackey¹, Eric Stevens^1,7, Dawn Y. Sumner^1,2
1Geology, University of California Davis, Davis, CA, USA
²W.M. Keck Center for Active Visualization in the Earth Sciences, U.C. Davis, Davis, CA, USA
³Computer Science, University of California Davis, Davis, CA, USA
⁴now at Retro Studios, Austin, TX, USA
⁵now at University of Texas, Austin, TX, USA
⁶now at Brown University, Providence, RI, USA
⁷now at University of Minnesota, Minneapolis, MN, USA

Abstract

1. Introduction: The W. M. Keck Center for Active Visualization in the Earth Sciences (KeckCAVES) provides a collaborative environment using interactive visualization methods for interpretation and modeling of 3D Earth structure and processes.  There are three facets to this environment: people engaged in tight collaboration across the disciplines of geoscience and computer science; software to analyze point, surface, and volume data to address geoscience problems; and hardware, including a 4-sided Cave Automatic Virtual Environment (CAVE) and low-cost immersive visualization environments based on 3D TVs.  Key contributions of the KeckCAVES project to 3D structural geology are the software suites and the research and development workflows that we have developed by partnering geologists and computer scientists. Our collaboration has led to development of the open-source Vrui (Virtual Reality User Interface) software platform (Kreylos, 2008; Kreylos et al., 2003), the basis for three applications for interactive research and teaching of 3D geologic structure: 3D Visualizer for analysis of volume data (Billen et al., 2008), Crusta, a virtual-globe that renders arbitrarily large terrain data and enables remote geologic mapping (Bernardin et al., 2011), and LiDAR Viewer for analysis of lidar point clouds (Kreylos et al., 2008; Kreylos et al., in press). The Vrui toolkit and its applications run on Unix/Linux and Macintosh platforms and are independent of specific visualization hardware environments.  The software tools employ highly efficient approaches to data handling and rendering, such as hierarchical data structures and view-dependent, multi-resolution, out-of-core rendering.  This means that the applications only load those parts of a dataset that are within a user's current view and then only render the visible parts of the data at the resolution supported by the hardware.  As a result, datasets can be visualized in real time that are many times larger than the computer's main memory. Our approach exploits the human capacity for visually identifying meaningful patterns embedded in noisy data by using immersive, interactive data visualization in real time.  Here, immersive means that users see, and are surrounded by, a virtual scaled model of the data; interactive means that users immediately see the results of operations they perform on those data without waiting for the view to refresh; and real time means the display updates at the minimum rate of 48 frames/second needed to support interactive stereoscopic rendering (Kreylos et al., 2003)

2. How We Work: We start by creating an environment that allows students, postdocs, and faculty in computer science and geoscience to self-organize into dynamic research teams focused on using advanced scientific visualization to address specific research problems.  These teams work together intensively, often meeting daily to weekly, to achieve research goals through use and development of new software tools.  We develop software iteratively as a tight collaboration between developers and users by focusing on rapidly creating working solutions to the most essential needs first, rather than seeking to develop more complex global solutions.  We also provide formal opportunities for training and career development at all levels, and outreach to K-12 students and teachers.

3. Examples/Case Studies:
3a. 3D Visualizer: 3D Visualizer is an immersive application for real-time interactive visualization of gridded volumetric scalar, vector, and/or tensor data, such as tomographic scans or convection simulations.  3D Visualizer provides a range of visualization methods, including arbitrarily-oriented color-mapped slices, isosurfaces to map the extent of a single value, or volume rendering to assign every voxel a value-dependent color and opacity. Colors and opacities can then be edited in real time to reveal features and reduce the visual clutter of noise.  Vector data, such as the output of flow simulations, can be rendered using tracer particles, arrows, and streamlines.  These capabilities have enabled researchers to investigate microbialite morphology in modern and fossil samples, allowing quantitative characterization of the dips of surfaces, the sizes of features, and various microbial growth features (Stevens et al. 2011; http://goo.gl/OzzXv).  These capabilities have also enabled researchers to assemble geophysical data into a large-scale structural model of the Aleutian subduction zone, for use as the basis of a geodynamics model (Jadamec and Billen, 2010; http://goo.gl/Otwe8).

3b. LiDAR Viewer: LiDAR Viewer is an immersive application for real-time interactive visualization of multi-billion-point LiDAR point clouds without sub-sampling or reduction in data size (http://goo.gl/dH1Q8). In LiDAR Viewer, point data are preprocessed and organized into an octree structure to provide a multi-scale representation and efficient localized-point-data retrieval, which enables seamless real-time visualization of arbitrarily large data sets.  It supports real-time navigation; real-time interactive illumination of the point cloud using multiple light sources at arbitrary positions; point selection and extraction for further processing; fitting of geometric primitives (line, plane, sphere, cylinder) to selected points; visualization of point distances from a plane; and extraction of profile curves. This 3-D interactive analysis of LiDAR data permits virtual field investigations, from which a geologist may rapidly compile structural measurements, taking advantage of perspectives and scale available only within a virtual space. In a study of the 2010 El Mayor-Cucapah earthquake surface rupture, Gold et al. (In Press) collected a dense data set of offset landforms from centimeter-resolution terrestrial lidar point clouds of fault scarps (http://goo.gl/t4QjH). Over ten times the density of unique offset features were detected versus those found during post-earthquake fieldwork. Virtual fieldwork mimicked that of the field geologist; offset features were projected to the fault plane to define ends of a slip vector. In the virtual space, features could be inspected from any view angle, including physically impossible vantages such as from underneath the ground surface. By measuring each offset feature multiple times, Gold et al. (In Press) empirically constrained the error distribution of each measurement -- a level of analysis that is infeasible in the field.

3c. Crusta: Crusta is an immersive virtual globe that accurately renders, in real time, DEM and image data with both large coverage (e.g., whole-earth) and high-resolution (e.g., sub-meter) while minimizing the performance requirements of the computer's graphics subsystem (http://goo.gl/4zUdj).  Crusta employs a polyhedral subdivision of the sphere to provide near-uniform and isotropic resolution everywhere on the sphere.  Crusta enables data exploration and feature discovery via interactive surface shading, real-time vertical exaggeration, and interactive manipulation of the texture color palette.  Topographic features, such as fault scarps or landforms displaced by coseismic surface rupture, can be documented by directly mapping on the virtual landscape using standard GIS formats (shapefiles, kml, etc.).  Forte et al. (2010) used a precursor to Crusta called the Real-time Interactive Mapping System (Bernardin et al., 2006), to remotely investigate late Cenozoic deformation within the Kura fold-thrust belt along the southeastern margin of the Greater Caucasus Mountains in Azerbaijan and Georgia by estimating bedding attitudes and mapping detailed structural geometries.  They then integrated these structural data with unit ages from published maps to compile a regional geologic map and construct two balanced cross sections across the thrust belt. The study suggests that the Kura fold-thrust belt is a first order structural system within the Arabia-Eurasia collision and has accommodated upwards of 50% of total Arabia-Eurasia convergence since its late Pliocene initiation.

3d. Teaching: Integration of 3D Visualizer and LiDAR Viewer into the undergraduate structural geology lab program introduces students to 3D visualization techniques and builds their 3D cognitive skills in preparation for field exercises. With 3D Visualizer students model the exposure of geologic structures across landscapes. They combine geologic structures and landscapes, discover how exposures are dictated by the topography, and learn to predict outcrop geometries based on structural data. Using LiDAR Viewer, students collect virtual field measurements from the Raplee Ridge anticline of southeast Utah. The structure of this asymmetric, doubly plunging fold is spectacularly revealed by the erosion of the San Juan River and tributary channels through tilted Paleozoic strata. Students learn to collect strike and dip measurements by fitting planes to bedding. The concept of three-point problems to derive good plane fits naturally follows from student exploration and measurement of the lidar data set. By the end of these exercises, students have learned to collect data from a variety of exposure geometries, and have gained valuable skill in visualizing geologic structures.

Students in a graduate-level advanced mapping class have used Crusta as part of a self-designed, discovery-based final exercise in which they use Crusta to remotely map bedrock structures with topographic expression, active structures, tectonic landforms, and surficial deposits.  Students learn how to both visually synthesize diverse geological datasets and make and record observations of the detailed topographic features that attest to active deformation of the landscape.

4. Summary and Conclusion: Geoscientists face growing challenges visualizing, measuring, and understanding very large and diverse data sets. These challenges are the fodder for innovation at intersections with computer science and cognitive science. The approach we take is to foster an open, cross-disciplinary collaboration that develops research problem-focused software solutions: 3D Visualizer for volume data, LiDAR Viewer for point-cloud data, and Crusta for structural mapping. These tools are available as free, open source software (www.keckcaves.org).

References:
Bernardin, T., Cowgill, E., Gold, R.D., Hamann, B., Kreylos, O., Scmitt, A., 2006, Interactive mapping on 3-D terrain models; Geochemistry Geophysics Geosystems, v. 7. doi: 10.1029/2006GC001335.
Bernardin, T., Cowgill, E., Kreylos, O., Bowles, C., Gold, P., Hamann, B., Kellogg, L.H., 2011, Crusta: A new virtual globe for real-time visualization of sub-meter digital topography at planetary scales: Computers & Geosciences, v. 37, p. 75-85.
Billen, M.I., Kreylos, O., Hamman, B., Jadamec, M., Kellogg, L.H., Staadt, O., Sumner, D.Y., 2008 ,A Geoscientist's Perspective on Immersive 3D Data Visualization: Computers and Geosciences, v. 34,  p. 1056-1072, doi: 10.1016/j.cageo.2007.11.009.
Forte, A.M., Cowgill, E., Bernardin, T., Kreylos, O., Hamann, B., 2010, Late Cenozoic deformation of the Kura fold-thrust belt, southern Greater Caucasus; Geological Society of America Bulletin, v. 122, p. 465-486.
Gold, P.O., Oskin, M.E., Elliot, A.J., Hinojosa-Corona, A., Taylor, M.H., Kreylos, O., Cowgill, E., In Press, Coseismic slip variation assessed from terrestrial lidar scans of the El Mayor-Cucapah surface rupture; Eartha and Planetary Science Letters.
Kreylos, O., Bethel, E.W., Ligocki, T.J. and Hamann, B., 2003, Virtual-Reality Based Interactive Exploration of Multiresolution Data, in: Farin, G., Hamann, B., and Hagen, H., eds., "Hierarchical and Geometrical Methods in Scientific Visualization," Springer-Verlag, Heidelberg, Germany, p. 205-224.
Kreylos, O., Bawden, G.W., and Kellogg, L.H., 2008, Immersive Visualization and Analysis of LiDAR Data, in: Bebis, G., et al., "ISVC 2008, Part I," LNCS 5358, Springer-Verlag Berlin Heidelberg, p. 846-855.
Kreylos, O., Cowgill, E.S., Oskin, M.E., Gold, P.O., Elliott, A.J., Kellogg, L.H., In Press, Point-Based Computing on Scanned Terrain with LiDAR Viewer; Geosphere.
Jadamec, M., Billen, M.I., 2010, Reconciling surface plate motions with rapid three-dimensional mantle flow around a slab edge; Nature, v. 465, p. 338-341, doi: 10.1038/nature09053.
Stevens, E.W., Sumner, D.Y., Harwood, C.L., Crutchfield, J.P., Hamann, B., Kreylos, O., Puckett, E. and Senge, P., 2011, Understanding microbialite morphology using a comprehensive suite of three dimensional analysis tools; Astrobiology, v. 11, p. 509-518. doi: 10.1089/ast.2010.0560.

AAPG Search and Discovery Article #120140© 2014 AAPG Hedberg Conference 3D Structural Geologic Interpretation: Earth, Mind and Machine, June 23-27, 2013, Reno, Nevada