In seismic exploration, surface waves are still an important topic, mainly because they hide the information from the depth recovered using body waves. On the other hand, experiment at laboratory scale are relatively little used for studies implying high spatial density. In this context, this thesis allowed to validate an automated laboratory environment dedicated to surface waves study, especially in presence of a surface and body waves mix comparable to the one who exist on the field. In a second step, surface and body waves have been efficiently seprated using array processing. After separation, it became possible to monitor waves arrival time in presence of medium variation at depth as well as at the surface, as in the case of oil reservoir monitoring (4D). Un full 4D study has been realized, allowing not only arrival-time variations monitoring, but also amplitude and arrival and departure angles. A methode to compensate for the spurious variations from the surface has been proposed. This last study has been completed by another one using field data. Then, the adaptation of High resolution algorithm allowed to consider the case of waves with close departure and arrival angles. Finally, taking advantage of the measurment high spatial density, a comparison of two designs - one theoreticaly ideal but not realistic and the other economically viable but sub-optimal - allowed to consider the issue ot the diffracted waves that disturb current acquisition. A new filtering solution has benne proposed for the second design.
