Noise Analysis and Image Restoration for Optical Sparse Aperture Systems

Optical sparse aperture imaging systems can capture the same resolution as conventional filled aperture with a reduction in size and weight, but their greatly reduced modulation transfer function (MTF) cause significant blurring and loss of contrast in the collected imagery. The presence of noise degrades the imagery further. Image restoration algorithms can correct the blurring completely when the signal to noise ratio (SNR) is high, but only partially when the SNR is low. We analyze the noise model of optical sparse aperture imaging systems and indicate that the total noise can be approximated as Gaussian noise. Noise amplification problem is also discussed based on traditional Wiener filter. An improved Wiener filter with Gaussian window is present, which firstly constructs a new degraded image using modified Lagrangepsilas interpolation technique, then estimates the SNR without trail and error approach, lastly restores the new degraded image. Results of a computer study for Golay6 sparse aperture imaging systems show the efficiency of the new algorithm.