PhD-project title: Comparative Visual Analysis Methods for Integrated Modelled and Measured Data
When comparing measured and modelled data, an intuitive way to understand similarities and extract dissimilarities is to use interactive 3D scientific visualization methods. A visual highlighting of regions with low correlation between measured and modelled data hints to problems in the underlying model. Extracting features like isocontours from both data collections shows the degree of concurrence in regions of interest. Using 3D visualizations at interactive frame rates while modifying certain parameters in the model, allows for an interactive assimilation of the model to the measured data: The comparative visualization gets updated while interactively tuning the parameters. For computing the correlation of measured and modelled data, one should take into account the uncertainty of interpolated data values.
Many different techniques (e.g. direct volume rendering, isosurface extraction, terrain rendering, flow visualization, etc.) may come to use for the scientific data visualization purposes: For measured data, the integration of data from different sources (paleo, seismic, satellite, etc.) needs to be taken care of. Sophisticated direct rendering techniques from scattered data may be applied or field reconstruction using scattered data interpolation techniques. Interpolation needs to take into account different scales in space and time, and possibly even constraints placed by obstacles. As many parameters are measured and modelled, the visualization techniques need to be able to visualize multi-variate data to see correlations between different parameters. A combination of classical scientific visualization techniques with abstract information visualization techniques may be beneficial. To deal with the large amount of data at interactive rates, a sophisticated data management should be applied, possibly using multiresolution methods, adaptive refinement, out-of-core techniques, and even parallel computations.
Start of doctoral thesis: 1st October 2011
Prof. Dr. Lars Linsen (JU), Prof. Dr. Reiner Schlitzer (AWI)
Committee Meetings: 13.07.12
Skills & interests:
- Software Development, including:
- Object-oriented analysis and design;
- Frameworks: Qt, MFC;
- Operating Systems: Linux, Windows;
- Version control (Subversion), Image processing, Sound processing;
Helmholtz Research School on Earth System Science, Annual Retreat 2012, Bremerhaven, “Deutsches Auswandererhaus“, Talk: Antonov, A et al., “Analysis of relevant fields in multifield data.“ 30 November, 2012.