A method for single image restoration based on the principal ergodic

We present a method to extract from a single image both object and point spread function using low contrast features of an extended field of view. Invoking the principal ergodic on stochastic turbulent phenomena, we show that the aberration parameters, characteristics of the earth's turbulence, can be recovered from multiple features within an isoplanatic patch. The ensemble statistics is replacing the spatial statistics of a single realization to derive an equivalent modulation transfer function and to apply usual deconvolution techniques such as Richardson-Lucy algorithms. The reliability of this postprocessing treatment has been tested on synthetic data, on solar granulation observations performed at La Lunette Jean Rosch du Pic du Midi, and during the event of the Venus transit at La Tour Solaire de Meudon.