INTERACTIVE IMMERSIVE VISUALIZATION OF GEOSCIENCE DATA

The geosciences are increasingly challenged to manage, process, visualize and interpret the large quantities of data generated by high-accuracy, high-resolution imaging and sensing technologies, or large-scale computer simulations of complex phenomena. We are developing and using interactive visualization software to view, interact with, and manipulate observed and/or simulated geophysical, geodynamical, and geologic data. The innovation of our approach is the highly effective use of human interaction in immersive 3D virtual reality (VR) environments. While immersive (head-tracked stereoscopic) visualization allows us to detect features in large and complex data more effectively, interactive tools substantially simplify the construction or manipulation of 3D shapes to isolate and identify those features, and to perform quantitative measurements of structures emerging from the data. We tailor our visualization approach to the specific scientific problems to take full advantage of each visualization method's strengths, using both 3D perception and interaction with data and ongoing simulations, to fully utilize the skills and training of geoscientists exploring their data in VR environments. In the process, we are developing a suite of tools that are adaptable to a broad range of scientific and engineering problems. We demonstrate our approach on several geophysical and geological datasets, including tripod-based and airborne LiDAR data, seismic tomography, geodynamics computational models, volumetric data from thin serial sections of Archean rock samples, and earthquake hypocenters. A LiDAR viewer (see Figure 1) allows a user to be fully immersed in point cloud data to assess data quality and to analyze complex targets that are hard to identify in standard, non-immersive, visualizations. Software mapping tools allow "virtual field studies" in regions that are otherwise inaccessible to human geologists. A visualization software for gridded volumetric data (see Figure 2) focuses on exploratory data analysis by supporting real-time extraction of derived shapes such as color-mapped slices, isosurfaces, or particle traces. Interactive measurement tools then allow scientists to quantify their observations. Our software is based on a "VR operating system" that can be used with standard computers and a range of immersive 3D environments (e.g., GeoWall, ImmersaDesk, CAVE). We have demonstrated that our approaches have clear advantages over other commonly used data analysis methods. Additional information about our work can be found at http://www.keckcaves.org.

Figure 1: Screen shot from the LiDAR viewer, showing a part of UC Davis campus (including the UC Davis water tower and the Mondavi Center for the Performing Arts). The image shows how a user can select features by using a "3D paint brush" (selected points are highlighted in green), and can quantify the position/orientation of features by, for example, extracting equations of best-fit planes (planes are visualized as green transparent rectangles).

Figure 2: Screen shot from the gridded volumetric visualizer, showing an early model of a subduction zone underneath Alaska. The subducting slab was isolated by an isosurface of viscosity, and structures in the surrounding mantle are visualized using color-mapped slices at user-defined positions and orientations.