Dependent component analysis for blind restoration of images degraded by turbulent atmosphere

Atmospheric turbulence is an inevitable problem in long-distance ground-based and space-based imaging. The optical effects of atmospheric turbulence arise from random inhomogeneities in the temperature distribution of the atmosphere. A consequence of these temperature inhomogeneities is non-stationary random distribution of the refraction index of the atmosphere [1]. Atmospheric turbulence can make distant objects being viewed through a sensor (e.g., a digital camera or video recorder) to appear blurred. Also, the time-varying nature of the turbulence can make the appearance of objects to wave in a slow quasi-periodic fashion. When a target is small and moving, its actual location becomes very difficult to estimate. This phenomenon greatly hinders accurate target detection, tracking, classification, and identification