Fast demosaicing technology for infrared focal-plane-array polarization imaging based on superpixel stokes vector...

Area of Science:

• Optics and Photonics
• Image Processing
• Computational Imaging

Background:

• Demosaicing is crucial for polarization filter arrays (PFAs) in division of focal plane (DoFP) imaging.
• Existing methods struggle with instantaneous field-of-view (IFoV) errors and low polarization reconstruction accuracy.

Purpose of the Study:

• To develop advanced demosaicing techniques for DoFP polarization imaging systems.
• To improve polarization reconstruction accuracy and address IFoV errors.

Main Methods:

• Introduced novel prior constraints based on PFA superpixel methodology, focusing on extinction ratio and non-uniformity.
• Developed a superpixel Stokes vector frequency reduction approach with inverse distance weighted (IDW) interpolation for reconstruction.
• Calibrated the imaging system to obtain prior parameters for enhanced demosaicing.

Main Results:

• The proposed method demonstrated comparable performance to state-of-the-art techniques.
• Achieved closer approximation to ground truth polarization parameters.
• Reduced computational complexity and showed suitability for embedded systems.

Conclusions:

• The novel demosaicing approach significantly enhances polarization reconstruction in DoFP systems.
• The method offers a practical solution with reduced complexity for real-world applications.
• This work advances PFA-based polarization imaging quality and efficiency.