The success of speckle interferometry in recovering high resolution information from atmospheric turbulence degraded images has renewed interest in the restoration of images recorded through a turbulent medium. There are a number of different approaches to image restoration which have been proposed. The effectiveness of each of these techniques is strongly dependent on the brightness and the angular extent of the object being observed. This paper will attempt to summarize the areas of application of such techniques, the current state-of-the-art in this field, and the expected performance range of the various techniques.

The concepts of speckle interferometry as developed by Labeyrie, and of speckle imagery as formulated by Knox and Thompson are analyzed for dependence on field-of-view size. The preliminary analysis, assuming isoplanatism rederives the results of Korff, and derives the result previously inferred by Knox and Thompson from computer simulation, that allowable spatial frequency separation for difference of phase shift determination must be less than r0/λ When the assumption of isoplanatism is dropped, results are obtained for the expected object power spectrum in speckle interferometry and for the expected bispectrum in speckle imagery, showing the dependence on angular spread for an object consisting of a pair of point sources. An angle, ϑ, is defined (in terms of an integral over the strength of turbulence distribution along the propagation path), which bounds the range within which there are no significant anisoplanatism effects. It is noted that the effect of anisoplanatism is not to attenuate the information bearing signal but rather to impose incorrect information on the signal. Thus anisoplanatism can result in incorrect conclusions with no indication that there is a problem.

The dependence of the signal to noise ratio on the exposure parameters in speckle interferometry is analysed. The choice of the optimum exposure parameters, which are found to depend on certain statistical functions of the image intensity, is discussed. Preliminary results of measurements of these statistical functions are given.

By adapting the principles of interferometry first developed by Michelson, we have built a simple incoherent interferometer which operates efficiently from 2 to 20 microns

The design and performance of this instrument will be discussed More than 12 stars have been at least partially resolved at one or more wavelengths and, in several cases, their sizes and shapes have been determined A number of objects in the solar system have also been studied and useful studies of extragalactic nuclei are now possible. With the development of the new Multiple Mirror Telescope both the resolution and the sensitivity of this instrument will soon be extended.

Some examples are given of types of ground motion which hs.ve been observed with seismic and geodetic instruments. In general, away from sources of man-made noise, the amplitudes of ground strains and tilts in the microseism band with periods of a few seconds normally are of the order of a few times 10-10. Strain and tilt tides give amplitudes of a few times 10-8, and long term average strain and tilt rates at relatively good sites may be roughly 10-7/yr. Superposed on such relatively long wavelength motions can be larger local pier motions, which are likely to be irregular. However, very high ground stability has been demonstrated at a few sites The limitations of ground motions on long term amplitude interferometer stability are serious if applications such as earth rotation and polar motion measurements are considered, but at shorter periods the environmental problems will be almost completely atmospheric.

The benefits and limitations of the application of high angular resolution interferometry to the study of binary stars are considered. The principle advantage of such observations is that they provide data complementary to parameters determined by other techniques and permit the direct determination of masses, angular diameters and distances to systems which would otherwise be inaccessible to such complete analysis. These values then directly lead to linear diameters, absolute magnitudes, absolute emergent fluxes and effective temperatures. The prospects for significantly increasing our knowledge of these fundamental astrophysical parameters are extremely promising. The applications of interferometry to surveys for new binary stars, to star clusters, and to wider visual binaries with the goal of detecting planetary perturbations are also briefly considered.

Some of the Galilean satellites show evidence of two distinct surface types, a high albedo frost and a dark, presumably silicate material. We are beginning a program to map the spatial distribution of these components with a Michel son spatial interferometer. This will be done both in and next to the 2.0µm frost absorption band. While the mapping program is just beginning, this investigation has already yielded new information on the wavelength dependence of seeing in the infrared.

A speckle interferometer is described which has been used to measure the separations and position angles of binary star systems. The present instrument uses an image intensifier to enable the enlarged stellar images to be recorded on cine-film. The exposure time is typically 8 msec and the camera can be operated at up to 20 frames per second. The instrument incorporates a means of calibrating the magnitude difference of the binary star systems.

A pair of achromatising lenses have been designed which will correct the radial chromatic dispersion in the speckle patterns and allow exposures to be made in “white” light. A system is under development which uses an image intensifier and a plumbicon television camera to acquire the data and a hard-wired autocorrelator to perform the analysis in real time. With this system the limiting magnitude of the technique should be significantly improved.

The stellar intensity interferometer at Narrabri Observatory was operated for ten years. Measurements were made of 32 single stars of spectral type 0 to F and were used to establish the first completely empirical temperature scale for hot stars. The application of a high-resolution interferometer to stellar astronomy was explored: observations were made of binary, rotating and emission-line stars, limb-darkening, polarized light and of the effects of atmospheric scintillation and of Cerenkov radiation from the night sky. This programme was completed in 1973. An analysis has been made of the astronomical programme and value to stellar astronomy of a much larger and more sensitive instrument. An intensity interferometer capable of reaching stars of magnitude +7.3 with a baseline of 2 km has been designed and its cost has been estimated.

We have developed a digital speckle camera for use on the University of Arizona 90-inch telescope. This camera uses a CID detector to provide photon locations in an image, or an analog image for brighter objects. We have used this system to observe Saturn's satellites Rhea and Iapetus. Using a correlation speckle technique, we have determined the angular diameter of these objects to be 1487 ± 40 km for Rhea, and 1200 ± 132 km for Iapetus.

The Astronomy Department of the University of Sydney is currently exploring the feasibility of building a Micheson stellar interferometer using modern techniques and detectors to surmount the atmospheric and stability problems that limited the development of the original Michelson interferometers. The main part of this programme is the design and construction of an 11 m fixed baseline prototype instrument to establish whether fringe visibility can be measured with an accuracy better than ±2% over a range of atmospheric seeing conditions and instrumental parameters. The design of the prototype instrument, its status and prospects are discussed.

Infrared Interferometry promises to be a useful astrometric technique. Preliminary measurements of the star α Orionis made with a heterodyne interferometer exhibit phase coherence over a period of at least 1000 seconds. The measurements were equivalent to a positional determination of 60 second accuracy every 5 seconds of integration.