Computational spectro-polarimetric imaging with resolution-preserving demosaicking.

Area of Science:

• Optics and Photonics
• Computational Imaging
• Remote Sensing

Background:

• Conventional imaging systems are often limited to monofunctional spectral or polarization measurements.
• Integrating spectral and polarization information is crucial for advanced imaging applications.
• Existing systems struggle to capture both spectral and polarization data simultaneously at high resolution.

Purpose of the Study:

• To develop an integrated spectro-polarimetric imaging system capable of simultaneous spectral and polarization data acquisition.
• To achieve high spatial resolution in reconstructed spectral-polarimetric images.
• To enable extraction of Stokes parameters, degree of linear polarization (DoLP), and angle of polarization (AoP).

Main Methods:

• Utilized a division of focal plane (DoFP) polarization camera combined with 16 broadband filters.
• Employed an ISTA-ResUNet framework for spectral reconstruction.
• Developed a resolution-preserving demosaicking algorithm for polarization image recovery.
• Reconstructed data across 31 spectral bands (400-700 nm) at four polarization angles (0°, 45°, 90°, 135°).

Main Results:

• Achieved full spatial resolution (2448 × 2048) for polarization camera data.
• Demonstrated high spectral fidelity (98.87%) in reconstructing a 24-color chart.
• Successfully reconstructed spectral and polarization information with exceptional performance in simulations and experiments.
• Enabled extraction of Stokes parameters (S₀, S₁, S₂), DoLP, and AoP.

Conclusions:

• The proposed spectro-polarimetric imaging system effectively integrates spectral and polarization measurements.
• The developed computational framework ensures high-resolution and accurate data reconstruction.
• This system offers a powerful tool for applications in remote sensing, material inspection, and biomedical imaging.