Autor: Imanol De Arriba Ruiz

Directores: Felix Perez Martinez / José María Muñoz Ferreras (Universidad de Alcalá de Henares)

Fecha de Exposición y Defensa: 19/12/2012

SAR (Synthetic Aperture Radar) and ISAR (Inverse SAR) techniques are coherent high-resolution radar systems, capable of providing a map of target radar cross section in the spatial domain of range and azimuth. The aim of both techniques consists in achieving a finer azimuth resolution by generating a synthetic aperture from the relative motion between radar and target.
Radar imaging complements the task of conventional optical and infrared systems, especially in adverse weather conditions. Conventional SAR and ISAR systems are designed for imaging targets with a direct line of sight of sensors. For this reason, they have a reduced performance in complex scenarios, such as in forests or urban environments, where multiple-bounce returns are superposed to direct-scatter echoes coming
from targets. They are known as ghost artifacts, since they obscure true targets and lead to poor visual quality, making target detection particularly difficult. The multipath mitigation problem for radar imaging is of theoretical and practical importance. In this Ph.D. thesis, Time Reversal (TR) concept is used as a new approach for SAR and ISAR to improve image visual quality by removing ghosting artifacts caused by multipath propagation (TR-SAR and TR-ISAR algorithms, respectively). Nevertheless, prior to applying these innovative multipath mitigation
techniques, it is needed to solve the geometric problemrelated to multipath. Focusing on the improvement of the performance of TR-ISAR, a series of advanced signal processing techniques are implemented before and after the time reversal-based stage (the central point of this thesis). The former ones (pre-processing techniques) are related to multilook averaging, time-frequency transforms and Radon transform, whereas the latter ones (post-processing techniques) are made up of a set of superresolution algorithms. In a nutshell, they all can be
seen as an added value to the TR concept, instead of being regarded as independent approaches. To sum up, the use of the designed time-reversal-based algorithm, along with
some of the proposed signal processing techniques, should not be circumvented if high-quality ISAR images are desired to be obtained in a dense-multipath environment. In fact, the resulting images may be useful for subsequent Automatic Target Recognition (ATR) schemes. As a proof-of-concept, both simulated and experimental data from high-resolution radars are used to verify the proposed methods.