Enhancing phase unwrapping by noisy pixels identifying criteria and iterative adaptive filtering

Area of Science:

• Optical metrology
• Image processing
• Signal processing

Background:

• Phase unwrapping is crucial for quantitative phase imaging.
• Noise in wrapped phase maps leads to significant errors and artifacts.
• Existing methods often struggle with noise, causing fake phase jumps.

Purpose of the Study:

• To develop a novel technique for accurate phase unwrapping in noisy conditions.
• To minimize phase jump errors caused by noise and phase residues.
• To enhance the robustness and accuracy of phase reconstruction.

Main Methods:

• A novel noisy pixels identifying criteria (NPIC) is proposed.
• NPIC leverages fringe pattern information to identify noise-contaminated pixels.
• A locally selective adaptive filtering strategy with dynamic window size and termination criterion is employed.
• The denoised phase map is unwrapped using the transport of intensity equation (TIE).

Main Results:

• The proposed method effectively identifies and filters noisy pixels, minimizing fake phase jumps.
• Locally selective adaptive filtering avoids blurring and detail loss associated with global filtering.
• The dynamic filter window and termination criterion ensure stability and efficiency.
• Superior phase reconstruction accuracy and robustness were confirmed through simulations and experiments.

Conclusions:

• The developed noisy pixels identifying criteria and iterative adaptive filtering significantly improve phase unwrapping.
• The technique demonstrates high performance even under complex and noisy measurement conditions.
• This method offers a robust solution for accurate phase reconstruction in optical metrology.