Real-time approaches for model-based PIV and visual fluid analysis

In this research project, approaches for the reliable reconstruction of flow fields from captured particle images and their visualization have been developed. One aspect has been on developing techniques that can generate a velocity field that is consistent with a selected physical fluid model. Therefore, we have introduced a model-based approach that integrates a priori knowledge of this model into the reconstruction process. Another aspect has been on the design of techniques that are capable of dealing with real-time constraints, and which thus have the potential to be used in combination with high-speed camera systems to interactively steer the reconstruction process. Programmable graphics hardware has been exploited as a co-processor for numerical computations to achieve interactivity, both for the reconstruction and visualization of generated fields. All these techniques have been verified in an experiment on living microorganisms. In the last phase of the project we have focused on the extension of the techniques towards the processing of 3D particle images and the visualization of the reconstructed flow fields.