This work presents the experimental results of a wavelet transform-based denoising algorithm to improve the signal-to-noise ratio (SNR) in Proton Sound Detectors (ProSDs). ProSDs are thermoacoustic detectors that exploit the acoustic signal generated by pulsed proton beams in energy absorbers to locate energy deposition with sub-millimeter precision, with interesting applications in beam monitoring during oncological hadrontherapy treatments. This work shows how the application of a denoising algorithm allows to improve the signal-to-noise ratio and reduce the extra-dose necessary to characterize the beam. The algorithm was validated with the thermoacoustic signal generated by a proton beam at 20 MeV, where it allowed to increase the SNR by 19 dB and decrease the dose necessary to characterize the beam by a factor of 80 compared to the state of the art.