One of the major constraints of measurements of atomic hydrogen densities using Two-Photon Absorption Laser Induced Fluorescence (TALIF) in most plasma and combustion environments is the determination of fluorescence decay times (τfluo H), especially when using nanosecond-lasers or slow acquisition systems. Therefore, it is necessary to identify the depopulation processes of the laser excited level in order to correctly estimate τfluo H. In this study, depopulation mechanisms of atomic hydrogen excited by two-photon absorption to the n=3 level (H(n=3)) have been investigated using a picosecond-laser excitation and acquisition of fluorescence by a streak camera, which allowed for direct measurement of τfluo H and hence, the atomic hydrogen densities, in a H2 microwave plasma operating in the pressure range 20 – 300 Pa. By combining these measurements with a detailed H(n=3) collisional radiative depopulation model, it was found that full mixing amongst the H(n=3) sub-levels occurs in our discharge conditions, even at a pressure as low as 20 Pa. Moreover, it is also seen that the Lyman β line is only partially trapped, as its escape factor Λ31 decreases from 0.94 – 0.98 down to 0.58 – 0.86 while the measured atomic hydrogen density rises from 8±5×1019 m-3 to 9±6×1020 m-3. This means that the radiative decay rate of H(n=3) atoms varies with pressure and the classical Stern-Volmer method used to determine the quenching cross-section of excited H(n=3) in collisions with H2 molecules, σQ H(n=3)/H2, is not valid for our measurements. We used two different physics-based approaches, and show that the quenching cross-section σQ H(n=3)/H2 lies in the range 90 – 106×10-20 m2, which can be averaged as 98±8×10-20 m2. This substantially improved estimation of σQ H(n=3)/H2 obtained in this work will be useful for the accurate estimation of H(n=3) fluorescence decay times and therefore the atomic hydrogen densities.