In a few years from now the LBT (Strittmatter
1988; McCarthy et al. 1988)
will be available for optical interferometry. The LBT will consist of two
8.4 m telescopes on a common mount with the mirrors only 14.4 m apart from
center to center. Therefore, excellent uv-coverage can be obtained. The
largest baseline is 22.8 m and the angular resolution 1.22
of this
interferometer is 6.1 milli-arcsec at 
nm, i.e., 22.8 m/2.4 m =
9.5 times higher than the resolution of the HST at the same wavelength.
At 
nm the speckles in the LBT interferograms have a
width of about 1.22
6.1 mas and a length (corresponding
to the 8.4 m mirror diameter) of about 16.5 mas. The number of speckles
is approximately equal to the number of turbulence cells in front of the
telescopes.
For example, for a Fried parameter 
cm (
0.35 arcsec seeing)
and an 8 m class telescope, the number of speckles is about 400 per
interferogram.
In this paper we present the theory and computer and laboratory simulations of interferometric speckle masking imaging with the LBT. Diffraction-limited images were reconstructed by a combination of a modified version of the speckle masking method (Weigelt 1977; Weigelt & Wirnitzer 1983; Lohmann et al. 1983; Reinheimer & Weigelt 1990; Reinheimer et al. 1993) and the building block method (Hofmann & Weigelt 1990; Hofmann & Weigelt 1993).
Figure 1: Time dependence of the LBT uv-coverage 
: the
two shaded areas are the uv-coverages at two different times
and 
Figure 2: Time dependence of the LBT uv-coverage and its influence on the
bispectrum coverage (see text)
Figure 3: Calculation of the bispectrum for complete and diluted uv-coverage.
In the case of the LBT and the shown instantaneous uv-coverage, only
bispectrum elements with u-vector, v-vector, and w-vector within the
dark shaded area can be measured. For example, the bispectrum element
corresponding to the shown vector triple cannot be measured during the shown
LBT uv-coverage or other times, but it can be measured with a single-dish
22 m telescope.