REM: A simplified revised entropy image reconstruction for photonics integrated interference imaging system

Abstract

The photonics integrated interference imaging system (PIIIS) is a recently developed far-field imaging approach aimed at compacting structures and acquiring high imaging performance. In this regard, to acquire higher imaging quality, several lens patterns have been developed, providing low-frequency intensive but high-frequency sparse measurements of the visibility function. However, sparse measurements at high frequency produce Gibbs-ringing artifacts in the image reconstruction, severely damaging the PIIIS imaging quality. In this study, a new data penalty for revised entropy is proposed for PIIIS, and a simple reconstruction algorithm using the Newton method is developed. A reconstruction experiment adopting the hierarchical multistage pattern is implemented on resolution board (USAF1951) targets with bright and black backgrounds. The outstanding improvement in the experimental results proves the feasibility of suppressing the ringing artifacts in the images reconstructed in PIIIS.