Inverse 2D obstacle scattering by adaptive iteration

The 2D inverse time harmonic electromagnetic scattering problem of reconstructing the starlike boundary /spl Lambda/ of an infinitely conducting obstacle from its far field scattering data is considered. An approach that employs weak a priori knowledge by choosing an auxiliary curve /spl Gamma/ inside the searched boundary /spl Lambda/ is used. Reconstructions are improved using an iteration scheme to adapt the internal curve /spl Gamma/ by exploiting information on the reconstruction /spl Lambda/ of the previous step. The adaptation algorithm yields significant improvements on /spl Lambda/, provided a reasonable first reconstruction may be obtained.