Efficient Polarization Demosaicking Via Low-Cost Edge-Aware and Inter-Channel Correlation

Abstract

Efficient and high-fidelity polarization demosaicking is critical for the industrial applications of division of focal plane (DoFP) polarization imaging systems. However, existing methods often struggle to balance speed, accuracy, and complexity. This study introduces a novel polarization demosaicking algorithm that interpolates DoFP images within a three-stage basic demosaicking framework. Our method incorporates a DoFP low-cost edge-aware technique (DLE) to guide the interpolation process. Furthermore, inter-channel correlation is used to calibrate the initial estimate in the polarization difference domain. The proposed algorithm is available in both lightweight and full versions, designed for different application requirements. Experiments on simulated and real DoFP images demonstrated that both versions achieve the highestt interpolation accuracy and speed, respectively, among existing interpolation-based algorithms and significantly enhanced visuals. The lightweight and full versions efficiently processed a 1024 × 1024 image on an AMD Ryzen 5600X CPU in 0.1402s and 0.2693s, respectively. Additionally, as our methods operate within a 5 × 5 window, parallel acceleration on graphics processing units (GPUs) or field-programmable gate arrays (FPGAs) is highly feasible.