A graphical data structure for complicated vector field properties and behavior
Dale McMullin · 2007
Proposing a standard for encoding vector field properties into a graphical data structure is potentially an important step in advancing progress in scientific analysis and visualization. This dissertation explores a potential definition of this standard, defines a library to implement it, and investigates several example implementations modeled from applications done in recent research. First, the use of vector field data in a variety of current research topics is presented as an introduction, and as the driver for this research. The formal thesis is stated. A conceptual model for the proposed data structure for vector field data is presented. Both scalar and directional fields are encoded to a color model through linear mapping, using the classical inverse square field as a guideline. The method for encoding and extracting data to and from an RGB color model is explored. Finally, the method for encoding 3D data as an array of 2D tiles in a single bitmap is discussed. The implementation of the data structure within an application, via an execution library is presented next. Additional vector properties for vector fields, in addition to the color model, is explored and justified. Finally, the order of operations and typical usage in an application execution environment is outlined and discussed. Following the presentation of structure and library, experimental implementations are discussed. The concepts of simplicity, abstraction, and exploitation are defined as presented in the thesis. Example implementations duplicated from areas of current research are deployed against the library, using variations of the proposed data structure presented earlier. These examples include implementations of complex field construction, collision detection, geometric modeling of computationally hard simulations, and modeling a complicated, large-scale event. Corollaries and potential areas of future work are revealed. These include alternatives to gaming (Boolean) property logic, memory tradeoffs, scalability benefits, and performance considerations. The final section of this dissertation presents a summary of conclusions.