Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz

AKPIK 5: Poster

AKPIK 5.3: Poster

Donnerstag, 12. März 2026, 15:00–16:30, P5

Stereovision-based angle and depth estimation for terahertz layer measurements — •Tim Arniko Meinhold^1,2, Dmytro Kharik¹, Joshua Hennig^1,2, Mirco Kutas^1,2, Jens Klier^1,2, Georg von Freymann^1,2, and Daniel Molter¹ — ¹Department for Materials Characterization and Testing, Fraunhofer Institute for Industrial Mathematics ITWM, 67663 Kaiserslautern, Germany — ²Department of Physics and Research Center OPTIMAS, RPTU Kaiserslautern-Landau, 67663 Kaiserslautern, Germany

Terahertz systems can measure layer thicknesses with high precision in a fully contactless, non-destructive manner. However, reliable results require accurate alignment of the beam perpendicular to the surface, as even small angular deviations can affect the measurements. Many optical approaches still struggle to combine precision and real-time capability on demanding surfaces. Meanwhile, advances in stereo imaging and machine-learning methods provide a framework to support stable, interpretable measurement conditions.

Here, we report a compact stereovision-based setup within a framework that includes neural network-based approaches. Our goal is a practical and reliable method that delivers precise depth and angular information, which are expected to improve terahertz real-time thickness measurements.

Keywords: Neural Networks; Optics; Terahertz; Position and Angle Calibration; Measurement system