Abstract

Particle Image Velocimetry (PIV) is a technique for optical measurement and analysis of particle flows. These flows are usually present in liquids, gases or can even consist of solid particles, like the ones created during the process of welding. Particle image velocimetry is a process that incorporates a wide spectrum of tasks and calculations to executed, which is why it is divided into several stages.

In this thesis, the challenges of implementing PIV on reconfigurable FPGA logic are elaborated. In addition, the requirements of each PIV stage, its structure and influ- ence on PIV are discussed. A special focus in this document is given to those parts of PIV which could be implemented in FPGA hardware. In the beginning, the theoreti- cal background of PIV is discussed that also includes a section about a new approach for extracting start and end-points of particles. The implementation part of this thesis contains besides the explanation of each stage's functionality, also the synthesis results of all projects, a visual step-through of stage results and the organization of project files. Finally, proposals for future improvements are discussed and a conclusion about the achieved PIV implementation is given at the end.

KEYWORDS: Particle Image Velocimetry, PIV, Detection of particles, Extraction of start and end points, Cross correlation, Fast Fourier Transformation, 1D-FFT, 2D-FFT, FFT CORE, Hardware implementation, Synthesis, Xilinx, Virtex 5, FPGA, VHDL