Presentation
3 October 2022 High accuracy dimensional microscopy through advanced modeling and machine learning
Phillip Manley, Ivan Sekulic
Author Affiliations +
Abstract
In the SiM4diM project we improve the measurement uncertainty of bidirectional optical measurements in industrial inspection to below a tenth of a micrometer. This will be achieved by combining highly accurate focal and afocal measurements with a robust model of the measured intensity of the structure in question. The inverse problem is then efficiently solved by applying a machine learning algorithm in the form of Bayesian optimization. We present a practical guide to modelling an optical system as well as the latest results of the SiM4diM project, showcasing improved edge detection over the state of the art.
Conference Presentation
© (2022) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Phillip Manley and Ivan Sekulic "High accuracy dimensional microscopy through advanced modeling and machine learning", Proc. SPIE 12216, Novel Optical Systems, Methods, and Applications XXV, 122160J (3 October 2022); https://doi.org/10.1117/12.2633288
Advertisement
Advertisement
KEYWORDS
Machine learning

Microscopy

Modeling

Microscopes

Numerical analysis

Optical imaging

Optical testing

Back to Top