Image Reconstruction from Sparse Irregular Intensity

A forward model-based algorithm for reconstructing images from Fourier domain measurements is described. The model accurately treats broadening of the Fourier domain sampling function resulting from finite aperture size and spectral bandwidth. The forward modeling approach also allows weighting of measurements according to their individual signal-to-noise ratios (SNRs), and injection of a priori expectations about the object such as its support and power spectral density. Image quality performance of a large telescope array is analyzed as a function of SNR and array scale. The approach is designed to enable joint processing of mixtures of intensity and amplitude interferometric measurements and conventional incoherent images, but the focus of this paper is limited to intensity interferometry. Basic structures of the object are seen to be recovered as SNR approaches 10 for unity coherence.