Digital Image Processing for Real-Time Correction of Atmospheric Turbulent Distortions

Algorithms for real-time processing of two-dimensional digital image during atmospheric path observations are considered. A procedure for measuring the turbulence parameter directly during observation is suggested. It can be applied both with the use of a reference object (a radial mire), and without it. An efficient high-speed correlation tracking algorithm with reference frame normalization is applied to correction of the image jitter caused by the atmosphere. Blurred image retrieval algorithms based on the inverse filtering of angular spatial frequencies are implemented using parallel computing procedures from the Intel® MKL and IPP libraries. The results of computer simulation of digital image blur and its correction are presented, as well as examples of processing experimental video frames along real atmospheric paths.