The Wavelet Transform is a tool which allows us to get information bot
h in the direct space and in the frequency space. We have developed an
algorithm based on the FFT with an isotropic wavelet. A multiresoluti
on deconvolution is performed for the interferometric images by using
the CLEAN method on the wavelet coefficients. An iterative reconstruct
ion assumes the existence of a solution which satisfies the positivity
and the limited support constraints, and which is compatible with the
measured visibilities. By this method, the images of two evolved star
s have been reconstructed from infrared speckle interferometry observa
tions.