Phase-shift estimation in sinusoidally illuminated images for lateral superresolution.

Sinusoidally patterned illumination has been used to obtain lateral superresolution and axial sectioning in images. In both of these techniques multiple images are taken with the object illuminated by a sinusoidal pattern, the phase of the sinusoidal illumination being shifted differently in each image. The knowledge of these phase shifts is critical for image reconstruction. We discuss a method to estimate this phase shift with no prior knowledge of the shifts. In postprocessing we estimate randomly introduced, unknown phase shifts and process the images to obtain a superresolved image. Results of computer simulations are shown.