Bio-inspired snapshot polarization-hyperspectral camera

Area of Science:

• Biomedical Optics
• Imaging Technology
• Computational Imaging

Background:

• Mantis shrimp possess highly advanced vision, capable of perceiving 16 color channels and polarization states.
• Existing imaging systems often struggle to capture both spectral and polarization information simultaneously and efficiently.

Purpose of the Study:

• To develop a snapshot polarization-hyperspectral camera (pHScam) inspired by mantis shrimp visual capabilities.
• To efficiently encode and reconstruct high-dimensional polar-spectral data for advanced imaging applications.

Main Methods:

• Designed a pHScam to capture linear polarization in four directions and spectral signatures across 21 bands, creating a 4D hypercube.
• Implemented bio-mimetic encoding: two-tier color encoding and pixelwise mosaic polarization encoding for 4D to 2D data compression.
• Utilized convex optimization with prior sparsity and system calibration for full-resolution data recovery.

Main Results:

• Successfully compressed the 4D polar-spectral data into a 2D mosaic raw pattern using novel encoding strategies.
• Demonstrated effective full-resolution recovery of the original 4D data through convex optimization algorithms.

Conclusions:

• The developed pHScam offers an efficient method for capturing and reconstructing rich polar-spectral information.
• Bio-mimetic encoding and convex optimization provide a powerful framework for handling complex hyperspectral and polarization data.