Single and double sequence of investigation domains in 2-d linear inversions applied to GPR Data

We focus on the choice of the investigation domains for a 2-D linear inversion algorithm based on the Born approximation. The application of an inversion algorithm to a wide B-scan requires the subdivision of the data into several observation domains and the partition of the buried scenario into different investigation domains, and this can be done in many ways, which raises the problem of examining the well advisedness of the adopted choice. In particular, here, we will propose the possibility of a double sequence of investigation domains. The results will also be compared with a migration algorithm. © 2004-2012 IEEE.