Adaptive Image Restoration Algorithm Based On Image Units Restructuring

Whether for scientific researching on celestial observation, or for national security considerations, the image quality of the space target observation is being more and more important. The proposing of Adaptive optics techniques greatly overcome the interference of atmospheric turbulence, which provides the possibility to obtain the higher quality image. However, because the adaptive optics technology is still not perfect, the adaptive images have to be restored with0 the digital image technology.This paper analyzes the atmospheric turbulence disturbance model, adaptive optics technology, observation model, elaborated common image restoration method, improved particle swarm algorithm and subjective and objective evaluation criteria.On this basis, this paper mainly studies three aspects:1. On the basis of fully explaining image quality evaluation methods, proposed an adaptive image restoration quality blind evaluation based on local variance variations to remove noise. Make statistical variance as the main evaluation methods focus on optimizing exclude artifacts and noise image quality evaluation of the interference, evaluation of adaptive image quality become more objectively and accurately.2. For lack of pre-adaptive image restoration algorithm, it propose a fragment reassembly method, in order to improve image quality, it splits multi-frame adaptive image into some image units which would be recombined by improved particle swarm algorithm to extract the local information of the high quality image, with which we achieve adaptive image pre-processing to improve the effect of restoration.3. To improve the RL and RL-IBD algorithm restoration, get more edge information, we proposed a TV regularization of RL-IBD image restoration algorithms by introducing the point spread function of dynamic support domain, TV regularization, asymmetry factor automatically update algorithm to improve image quality.