Iterative normalized correspondence ghost imaging

We propose a technique, which we call iterative normalized correspondence ghost imaging, to remove noise and improve the quality of traditional ghost imaging (GI). An iterative model based on correspondence imaging is established by assuming invariance of the background noise between successive measurements. Numerical simulation is used to determine the optimal parameters of the model, including the number of iterations required. Both simulation and experimental results reveal that the quality of the image reconstructed by this strategy is higher compared to that of traditional correspondence GI and normalized GI, and no more demanding. The signal-to-noise ratio is improved without requiring a priori knowledge of the target object. This approach represents another step forward towards real practical applications.