Interferometric system with deep learning algorithm to process two interferograms

Abstract

Methods, devices and systems for measuring surface roughness and surface shape of an object are described. An example interferometric system includes a collimator and a first and a second light sources with different spectral ranges and different coherence lengths. The system selectively allows light from one of the light sources to reach the collimator, and also includes a beamsplitter, and a Mirau type microscope having an objective lens, a plate with a central reflective spot and a beamsplitter plate to produce a reference beam and a test beam. An imaging lens receives the test and reference beams that form a plurality of interferograms. A neural network receives two of the interferograms for measuring the surface shape and another two interferograms for measuring the surface roughness of the object. The interferometric systems have a compact form, making them suitable for on-machine measurements and other applications.