Abstract

By averaging over the power spectra of many far-field speckle intensity patterns scattered by a laser-illuminated rough object, one can, after subtraction of a dc term, estimate the autocorrelation of an incoherent image of the object. Equivalently, by averaging the autocorrelations of the many speckle patterns, one can, after subtraction of a bias term, estimate the squared modulus of the Fourier transform of the incoherent image. Then one can reconstruct the incoherent image, or equivalently retrieve the Fourier phase, using a phase retrieval algorithm such as the iterative transform algorithm. Image reconstruction is facilitated by the non-negativity constraint that is valid for incoherent images.

© 1987 Optical Society of America