Learning-based single-shot long-range synthetic aperture Fourier ptychographic imaging with a camera array

In this Letter, we report a new long-range synthetic aperture Fourier ptychographic imaging technique, termed learning-based single-shot synthetic aperture imaging (LSS-SAI). LSS-SAI uses a camera array to record low-resolution intensity images corresponding to different non-overlapping spectral regions in parallel, which are synthesized to reconstruct a super-resolved high-quality image based on a physical model-based dual-regression deep neural network. Compared with conventional macroscopic Fourier ptychographic imaging, LSS-SAI overcomes the stringent requirement on a large amount of raw data with a high spectral overlapping ratio for high-resolution, high signal-to-noise imaging of reflective objects with diffuse surfaces, making single-shot long-range synthetic aperture imaging possible. Experimental results on rough reflective samples show that our approach can improve the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) by 10.56 dB and 0.26, respectively. We also demonstrate the single-shot ptychography capability of the proposed approach by the synthetic aperture imaging of a dynamic scene at a camera-limited speed (30 fps). To the best of our knowledge, this is the first demonstration of macroscopic Fourier ptychography to single-shot synthetic aperture imaging of dynamic events.