Reconstruction method of computational ghost imaging under atmospheric turbulence based on deep learning

Abstract

Ghost imaging, as an emerging imaging method, has great advantages in harsh environment with its off-object imaging characteristics. In this paper, we use a turbulence model based compressive sensing computational ghost imaging system to simulate atmospheric turbulence, analyze the effects of various factors on the imaging results, and recover the images under extreme turbulence conditions using conditional generation adversarial network, which can finally recover the images well. The simulation results show that the image reconstruction method proposed in this paper can recover the image well under the condition of very low sampling rate (1.56%).