Tomographic image reconstitution of objects from multi-incidence microwave exploration

To compete with other methods of diagnosis, microwave imagery has to be given in three dimensions with good accuracy in all directions. Experiments carried out with a single incident-beam illumination gave a good transverse accuracy, but the longitudinal one was not as good. To obtain a good image reconstitution, new computational simulations were carried out using many different incident angles to improve longitudinal accuracy. The results show that a choice must be adopted between the image quality and the required computing ability. Criteria given in the paper can help in this choice. Finally, we have shown that three-dimensional images can be obtained with good accuracy from a multi-incidence exploration.