Optical stellar interferometers have demonstrated milli-arcsecond resolution with few apertures spaced hundreds of meters apart. To obtain rich direct images, many apertures will be needed, for a better sampling of the incoming wavefront. The coheren
t imaging thus achievable improves the sensitivity with respect to the incoherent combination of successive fringed exposures. Efficient use of highly diluted apertures for coherent imaging can be done with pupil densification, a technique also called hypertelescope imaging. Although best done with adaptive phasing, concentrating most energy in a dominant interference peak for a rich direct image of a complex source, such imaging is also possible with random phase errors such as caused by turbulent seeing, using methods such as speckle imaging which uses several short exposure images to reconstruct the true image. We have simulated such observations using an aperture which changes through the night, as naturally happens on Earth with fixed grounded mirror elements, and find that reconstructed images of star clusters and extended objects are of high quality. As part of the study we also estimated the required photon levels for achieving a good signal to noise ratio using such a technique.
Optical interferometry has been successful at achieving milliarcsecond resolution on bright stars. Imaging performance can improve greatly by increasing the number of baselines, which has motivated proposals to build large (~ 100 m) optical interfero
meters with tens to hundreds of telescopes. It is also desirable to adaptively correct atmospheric turbulence to obtain direct phased images of astrophysical sources. When a natural guide star is not available, we investigate the feasibility of using a modified laser-guide-star technique that is suitable for large diluted apertures. The method consists of using sub-sets of apertures to create an array of artificial stars in the sodium layer and collecting back-scattered light with the same sub-apertures. We present some numerical and laboratory simulations that quantify the requirements and sensitivity of the technique.
In a previous paper, we discussed an original solution to improve the performances of coronagraphs by adding, in the optical scheme, an adaptive hologram removing most of the residual speckle starlight. In our simulations, the detection limit in th
e flux ratio between a host star and a very near planet (5 lambda/D) improves over a factor 1000 (resp. 10000) when equipped with a hologram for cases of wavefront bumpiness imperfections of lambda/20 (resp. lambda/100). We derive, in this paper, the transmission accuracy required on the hologram pixels to achieve such goals. We show that preliminary tests could be performed on the basis of existing technologies.
