End-to-end aperture layout optimization and image restoration with balancing quality and cutoff frequency in optical sparse aperture systems

Optical sparse aperture (OSA) systems achieve spatial resolution comparable to large monolithic telescopes while significantly reducing weight, size, and cost. The majority of decoupled optimization approaches for aperture layouts and image processing algorithms introduces design redundancy and struggle to simultaneously achieve high imaging quality and preserve the system's theoretical cutoff frequency. In this paper, we propose an improved joint optimization design method that co-optimizes aperture configurations and image restoration network. The novel multi-loss formulation reduces design redundancy while balancing both imaging quality and the cutoff frequency of the OSA system. The simulations validate the feasibility of the proposed method and demonstrate its robustness against residual piston errors ranging from 0 to 0.1λ. Finally, experimental validation confirms that the co-optimized 9-sub-aperture configuration with its corresponding image restoration network achieves a cutoff frequency equivalent to an 18-mm full aperture while significantly improving image contrast.