Optical shaping self-mixing interferometry with a neural network for displacement measurement

Abstract

Based on the characteristics of optical shaping self-mixing interference (SMI) and its perfect alignment with the input requirements of neural networks (NNs) for phase extraction, a novel, to our knowledge, displacement measurement method is proposed in this work. Optical shaping involves using a static Fabry–Perot cavity to map the periodic variations of optical frequency generated by SMI, achieving fringe multiplication, signal normalization, and enhancement for SMI optically. A NN trained on simulated data is used to directly extract the phase from the spectrum-mapped SMI signal. This measurement technology achieves a relative accuracy of 10-3 and advances the development of SMI.