Abstract

Following Michelson [1] original experiment, the beams from a stellar interferometer are generally combined in the image plane [2]. Interference fringes are observed when the images produced by each aperture are superimposed. The fringe spacing is set by the apparent angular distance between the apertures as seen from the image. It can be set to any value and is insensitive to image motion or guiding errors. A Fourier analysis of the image gives the energy at the fringe spatial frequency as an estimate for the square of the fringe visibility, filtering out the noise at all other spatial frequencies [3,4]. A drawback of this design is that the images do not always exactly coincide due to turbulence induced image motion or guiding errors thus randomly reducing the apparent fringe visibility and therefore decreasing the signal-to-noise ratio.

© 1986 Optical Society of America