Adaptive quadrature filters for multiple phase-stepping images

Area of Science:

• Optics and Image Processing
• Computational Imaging

Background:

• Phase recovery from fringe patterns is crucial in optical metrology.
• Existing methods struggle with high noise levels in fringe images.

Purpose of the Study:

• To develop a robust technique for local phase recovery from noisy, multi-frame phase-stepping fringe images.
• To improve the accuracy and reliability of phase reconstruction in challenging imaging conditions.

Main Methods:

• Utilized adaptive quadrature filters based on Bayesian estimation theory.
• Employed complex-valued Markov random fields as prior models for noise handling.
• Developed a novel technique for local phase recovery.

Main Results:

• Achieved accurate phase reconstructions even with extremely noisy fringe images.
• Demonstrated that the technique's performance is largely independent of the specific noise model for wideband noise.
• Showcased the effectiveness of Bayesian estimation and Markov random fields in phase recovery.

Conclusions:

• The proposed technique offers a significant advancement in robust phase recovery from noisy fringe data.
• This method provides reliable phase information essential for quantitative phase imaging and metrology applications.
• The adaptive filtering approach enhances the applicability of phase-stepping techniques in real-world scenarios with inherent noise.