Since the first detection of a Gravitational Wave, the LIGO-Virgo Collaboration has worked to improve the sensitivity of their detectors. This continuous effort paid off in the last scientific run, in which the collaboration detected an average of one gravitational wave per week and collected 74 candidates in less than one year. This result was also possible due to the Frequency Independent Squeezing (FIS) implementation, which improved the Virgo detection range for the coalescence between two Binary Neutron Start (BNS) of 5-8\%. However, this incredible result was dramatically limited by different technical issues, among which the most dangerous was the mismatch between the squeezed vacuum beam and the resonance mode of the cavities. The mismatch can be modelled as a simple optical loss in the first approximation. If the beam shape of squeezed vacuum does not match the resonance mode, part of its amplitude is lost and replaced with the incoherent vacuum. However, this modelisation is valid only in simple setups, e.g. if we study the effect inside a single resonance cavity or the transmission of a mode cleaner. In the case of a more complicated system, such as a gravitational wave interferometer, the squeezed vacuum amplitude rejected by the mismatch still travels inside the optical setup. This component accumulates an extra defined by the characteristics of the mismatch, and it can recouple into the main beam reducing the effect of the quantum noise reduction technique. This issue will become more critical in the implementation of the Frequency Dependent Squeezing. This technique is an upgrade of the Frequency Independent Squeezing one. The new setup will increase the complexity of the squeezed beam path. The characterisation of this degradation mechanism requires a dedicated wavefront sensing technique. In fact, the simpler approach based on studying the resonance peak of the cavity is not enough. This method can only estimate the total amount of the optical loss generated by the mismatch, but it cannot characterise the phase shift generated by the decoupling. Without this information is impossible to estimate how the mismatched squeezed vacuum is recoupled into the main beam, and this limits the possibility to foreseen the degradation of the Quantum Noise Reduction technique. For this reason, the Padova-Trento Group studied different techniques for characterising Mode Matching. In particular, we proposed implementing the Mode Converter technique developed by Syracuse University. This technique can fully characterise the mismatch of a spherical beam, and it can be the first approach to monitoring the mismatch. However, this method is not enough for the Frequency Dependent Squeezer source since it cannot detect the mismatch generated by the astigmatism of the incoming beam. In fact, the Frequency Dependent Squeezer Source case uses off-axis reflective telescopes to reduce the power losses generated by transmissive optics. This setup used curved mirrors that induce small astigmatic aberrations as a function of the beam incident angle. These aberrations are present by design, and the standard Mode Converter Technique will not detect them. To overcome this issue, I proposed an upgrade of the Mode Converter technique, which can extend the detection to this kind of aberration.