A guided wave-based structural health monitoring (SHM) system aims at determining the integrity of a wide variety of plate-like structures, including aircraft fuselages, pipes, tanks etc. It relies on a sparse array of piezoelectric transducers for guided waves (GWs) excitation and sensing. With a number of benefits, these waves are standing out among other methods as a promising method for the inspection of large structures. They can propagate on significant distances with small attenuation while being sensitive to surface and subsurface defects.This thesis presents studies conducted with the purpose of developing such a GWs-based SHM system that is capable of efficient defect detection, localization and sizing aeronautical plate-like structures made of aluminum and composite materials. Simulation and data-driven approaches are presented for determining principal characteristics of propagating GWs, namely modal group and phase velocities, 3D Green's functions etc. in structures of interest. They are then used for GWs signals processing in order to compute images representing the integrity of studied structures. This work also provides a comprehensive overview of DAS, MV and Excitelet defect imaging algorithms, determines their performance using statistical analysis of an extensive dataset of simulated guided waves imaging (GWI) results and proposes a method for sparse defect imaging.While defect detection and localization are straightforward from the image analysis, the defect sizing is a more complex problem due to its high dimensionality and non-linearity. It is demonstrated that this problem can be solved by means of machine learning methods, relying on an extensive database of simulated GWI results. Aforementioned defect imaging methods are baseline demanding. They are efficient under stationary operational conditions but vulnerable to environmental variations, especially to the temperature fluctuation.Finally, this work presents studies on the robustness of GWI methods against thermal effects, and a defect detection model capable of analyzing deteriorated GWI results is proposed. Different techniques for thermal effects compensation are reviewed, and improvements are proposed. Their effectiveness is validated for aluminum plates but further improvements are required to translate these techniques to composite plates.