Phase unwrapping algorithm based on a rank information filter.

Area of Science:

• Image Processing
• Signal Processing
• Computational Imaging

Background:

• Phase unwrapping is crucial for many imaging techniques, but noise in wrapped phase images complicates accurate retrieval of unambiguous phase information.
• Existing phase unwrapping algorithms often struggle with noise sensitivity, leading to errors and reduced accuracy in reconstructed phase data.

Purpose of the Study:

• To develop a robust phase unwrapping algorithm capable of retrieving unambiguous phase from noisy wrapped phase images.
• To enhance the reliability and accuracy of phase unwrapping in the presence of significant image noise.

Main Methods:

• A novel recursive phase unwrapping program is proposed, framed within a rank information filter for state estimation.
• A local phase gradient estimator utilizing the amended matrix pencil model (AMPM) provides essential phase gradient information.
• An efficient heap-sort-based path-following strategy guides the unwrapping process from high-quality to low-quality pixels.

Main Results:

• The proposed algorithm demonstrates effectiveness in retrieving unambiguous phase from noisy wrapped phase images.
• Validation using both synthetic and experimentally measured data confirms the robustness of the rank information filter-based approach.
• The method successfully obtains reliable phase unwrapping solutions even with substantial image noise.

Conclusions:

• The developed rank information filter-based phase unwrapping algorithm offers a robust solution for noisy phase data.
• The combination of AMPM for gradient estimation and heap-sort for path planning enhances unwrapping accuracy and efficiency.
• This approach significantly improves the quality of unwrapped phase images in challenging noisy conditions.