High Resolution Image Reconstruction from Images Degraded by Heavy Atmospheric Turbulence

摘要

We describe an image restored algorithm for images degraded by heavy atmospheric turbulence acquired through modified deconvolution, which significantly selects the input sequence, beginning with some frames randomly, and gradually increasing the observed frames to generate satisfactory reconstructions. Firstly, the error function is constructed by a frequency ratio constraint in the low-frequency area to preserve the signal and suppress noise in the high-frequency area, then the image and PSFs are obtained by an iterative recursive projection scaled-gradient algorithm beginning with some frames randomly. The simulation employs phase screens to simulate the images from ground-based telescope systems that have been degraded by heavy atmospheric turbulence, which demonstrate the algorithm to be effective for such degraded images and especially to suit cases where some frames have been lost from the original observed images sequence.