An efficient iterative super-resolution technology for coded aperture imaging

In this paper, we employ coded aperture imaging (CAI), an emerging computational technology that captures 4D light-field information to realize pixel super-resolution imaging via post-processing. Our CAI experimental setup is built based on 4f delay system with reflective optical path structure, where a programmable LCOS spatial light modulator is integrated at the Fourier plane to implement high-resolution high-contrast aperture coding, without requiring specialized hardware or any moving parts. In addition, we propose an iterative super-solution reconstruction algorithm based on aperture coding, optical fields manipulation and compressed sensing. First, we establish an accurate mathematical model for the OTF of coded aperture system and pixel binning process. Then, based on a series of low-resolution intensity image, we computationally reconstruct the high-resolution image with the convex projection iterative algorithm. The effectiveness of this algorithm is demonstrated with both simulation and experimental results. Due to its flexibility and simplicity, this technology can break physical limitations of the detectors' resolution to one that is solvable through computation, rendering it a promising tool in public security, military survey, medical science and many other fields.