Abstract
Highlights This is the first time that high-precision measurement of large-aperture optical components has been realized by capturing single-frame interferograms without requiring a phase shifter. We integrate deep learning algorithms with interferometry methods to simulate the interferometric process in a deep learning framework. This approach mitigates environmental noise effects on measurement accuracy, eliminates phase shifters, and enables dynamic surface profile measurements of large-aperture optical components. Compared with traditional phase-shifting methods, this approach achieves a 48-fold time reduction while improving measurement efficiency and stability. We demonstrate a highly efficient dynamic measurement method that achieves comparable accuracy without ZYGO interferometers’ ultra-stable environment requirements.
Original language | English |
---|---|
Article number | 055601 |
Journal | International Journal of Extreme Manufacturing |
Volume | 7 |
Issue number | 5 |
DOIs | |
Publication status | Published - 1 Oct 2025 |
Keywords
- deep learning
- dynamic high-precision measurement
- large-aperture optical components
- single-frame interferometric
- surface profile