Eveline J. Ayme-Bellegarda
IBM T. J. Watson Res. Ctr., P.O. Box 704, Yorktown Heights, NY 10598
Tarek M. Habashy
Schlumberger-Doll Res., Ridgefield, CT 06877-4108
This work is concerned with the imaging and quantitative characterization of objects embedded in inhomogeneous material, using acoustic or elastic waves as the interrogating sources. Such problems are recurrent in many areas such as, for instance, geophysics, oceanography, and medical imaging. The interest here hinges on the determination of high contrast channels embedded in multilayered elastic structures. A general framework is presented, based on vector integral equations. In addition to being suitable for both the forward and inverse problem, this framework is appropriate for any geometry and transmitter--receiver characteristics. Tractability issues are discussed, leading to the consideration of the Born approximation for inhomogeneous elastic background. To go beyond the inherent limitations of this approximation, an iterative inversion scheme is introduced. The method is illustrated in the case of a 2-D problem. First, the forward model is validated against an independent means based on a finite-difference approach. Then, constrained least-square reconstruction is performed, aiming at recovering the density distribution of a 2-D fluid-filled object buried in high contrast layers.