Addressing dynamics in radio frequency tomography

Abstract

Tomographic imaging of a dynamic scene is addressed for the first time in this article. The effect of dynamics of the imaged object is studied for some deformation scenarios using the conventional truncated singular value decomposition (TSVD) method, which is applied in numerous static inverse scattering problems, together with variational data assimilation (DA) methods including three-dimensional variational (3D-VAR) and four-dimensional variational (4D-VAR) techniques. Results show that the DA frameworks outperform the TSVD method, which assumes the scene static, especially for objects with higher contrast. Also, it is seen that applying the conventional TSVD method makes the image blurred and distorts it for high-speed dynamic changes, particularly when the contrast is increased. However, the proposed DA methods seem to be robust against large deformation distortions and can still detect the object.