[Back]

L. Fritz:
"Interactive Exploration and Quantification of Industrial CT Data";
Supervisor: E. Gröller, M. Hadwiger; Institut für Computergraphik und Algorithmen, TU Wien, 2009; final examination: 2009-01.

Non-destructive testing (NDT) is a key aspect of present day engineering and development which examines the internal structures of industrial components such as machine parts, pipes and ropes without destroying them. Industrial pieces require critical inspection before they are assembled into a finished product in order to ensure safety, stability, and usefulness of the finished object. Therefore, the goal of this thesis is to explore industrial Computed Tomography (CT) volumes, with the goal to facilitate the whole quantification approach of the components at hand by bridging the gap between visualization on the one hand, and interactive quantification of features or defects on the other one. The standard approach for defect detection in industrial CT builds on region growing, which requires manually tuning parameters such as target ranges for density and size, variance, as well as sometimes also the specification of seed points. To circumvent repeating the whole process if the region growing results are not satisfactory, the method presented in this thesis allows interactive exploration of the parameter space. The exploration process is completely separated from region growing in an unattended pre-processing stage where the seeds are set automatically. The pre-computation results in a feature volume that tracks a feature size curve for each voxel over time, which is identified with the main region growing parameter such as variance. Additionally, a novel 3D transfer function domain over (density, feature size, time) is presented which allows for interactive exploration of feature classes. Features and feature size curves can also be explored individually, which helps with transfer function specification and allows coloring individual features and disabling features resulting from CT artifacts. Based on the classification obtained through exploration, the classified features can be quantified immediately. The visualization and quantification results of this thesis are demonstrated on different real-world industrial CT data sets.