Dominik Hofer,
"Contributions to optical measurement and image processing methods"
, 9-2021
Original Titel:
Contributions to optical measurement and image processing methods
Sprache des Titels:
Englisch
Original Kurzfassung:
Five peer-reviewed journal articles are presented in this dissertation. The topics span from the numerical approximation of the Huygens-Fresnel integral, over the measurement of currents in printed circuit boards (PCB) or through cast molten steel in a continuous casting plant, to the flow field of small electronics cooling fans. Despite the obvious differences, there is a common factor which ties these publications together: they all rely on image processing and optical methods in instrumentation and measurement. All points on a wavefront are sources of new spherical waves and their superposition is propagating this wavefront - this principle was solved numerically in `A numerical approximation of the Huygens-Fresnel integral' to simulate laser speckles. In `2D Current Distributions and Their Magnetic Field' a similar concept - the Biot-Savart law - was employed to estimate currents flowing through the traces of a PCB. `Image Processing for Calibrating a Coordinate Measurement Set-Up' deals with the calibration of a continuous steel casting simulator's sensor positioning system by means of optical methods and shows first results of the simulated flow measurements. In `High-resolution 2.5D Particle Image Velocimetry Measurements of the Flow Fields Generated by Small Fans' a new measurement method, a variant of PIV, was developed and tested. The 3D flow fields of a small axial and a centrifugal fan were measured and analyzed with image processing methods. The very same centrifugal fan was also the subject of the more detailed `Particle Image Velocimetry and Constant Temperature Anemometer Measurements of the Jet Produced by a Centrifugal Fan,' where the potential of the new measurement procedure was further demonstrated. In all the presented articles the advantages of the methods used become evident: they are non-invasive, return clear (graphic) results, and feature extraordinary precision, due to the quality of modern optics, cameras and algorithms.