The Influence of Noise on the Restoration of Diffraction-Limited Images

Abstract

In the smaller-aperture imaging system for aerospace application, the diffraction degradation in the image resulted from the finite-size aperture of an optical system is troublesome. A great deal of effort has been made to restore the degraded image. Unfortunately, the process of restoration is an ill-conditioned problem, especially for the images with severely noise. The paper presents an analysis of the influence of various noise including salt-and-pepper noise, Gaussian noise and periodic noise on the restoration of diffraction-limited images, with the method of Wiener filtering, constrained adaptive restoration algorithm and iterative blind restoration algorithm. It is proved that the iterative blind restoration algorithm is superior to the others for the image with Gaussian noise and salt-and-pepper noise, though the resistibility to salt-and-pepper noise of the three approaches is all weak. Wiener filtering is effective for eliminating the periodic noise while the other two is invalid, but the restored image is not sharp enough. It is suggested to remove salt-and-pepper noise with median-filter, and periodic noise with frequency domain band-reject filter before the restoration.