Microsoft Word - icassp12_shen_rev_120107_final4.docx

In this paper, we consider ultrasound imaging of flaws in a metallic alloy where the presence of strong bottom surface reflection and other interference signals constitutes a challenging problem. A subspace-based approach is developed for removing, or significantly reducing, bottom surface reflections to enhance ultrasound imaging. In constructing the surface reflection, or clutter, subspace, we account for rough surface scatterings which, due to various possible propagation time delays between the transmitter and receiver, expand the subspace dimension beyond that corresponding to ideal propagation. We also estimate and compensate, using signal correlation methods, for changes in the same time delays due to imperfect sensor displacements on the top surface of the alloy. Experimental results show that substantial clutter suppression can be achieved with negligible effects to the flaw signals. Index Terms — Ultrasound imaging, clutter suppression, subspace projection