Intelligent Automation, Inc. — Department of Defense SBIR Phase I: DTRA192-004

Description

Intelligent Automation, Inc. (IAI) proposes the application of a powerful imaging technique, based on the multiphase level set framework, for retrieving shapes, locations and dielectric properties of unknown objects within the building walls. To our knowledge, no other technique can simultaneously retrieves the object shapes, locations and their dielectric properties using measurements collected at one side of the wall. These tools are very beneficial for classification purposes, since many detection and classification algorithms rely on object shapes and topological features. Using the level set framework, the amount of a priori information from the operator is minimized. The problem is solved using an optimization approach, minimizing the error between measurement data and simulated ones. A frequency hopping technique is used during the inversion. Lower frequencies are used to find objects locations and their general appearances and higher frequencies are used to retrieve finer details within the wall. The transmitting and receiving antennas can be placed at far distances from the wall. The walls are assumed to be inhomogeneous. Our method shows excellent performance against noisy data. With the application of high performance computing tools, we expect to finish the calculations in a few seconds.