Abstract: The vibration of the wind tunnel testing model under the flow pulsation leads to the inclusion of the model movement velocity in the fluorescent oil film path velocity directly computed by optical flow method, decreasing the accuracy of surface friction measurement. Therefore, a velocity measurement for fluorescent oil film on a testing model surface is presented. The method uses the background texture (such as artificial grid lines or other typical features) as benchmark to obtain the map matrix (geometric posture) based on the image correlation method and the discretion match for the adjacent timing images. The global optimization equation is also derived based on the continuity of the model movement, and then the movements of the testing model and the fluorescent oil film are decoupled. The simulation of the fluorescent oil film movement for a Oseen vortex pair has demonstrated that the maximum relative velocity error along Oseen vortex cores is 4.1% for the given parameters of translation and rotation. It has only 0.6% increment compared with that of the optical flow method tackling no translation and rotation images. The measurement of fluorescent oil film on a cavity testing model in 2 m scale high-speed wind tunnels has further showed that not only the measured flow phenomenon is correct, but also it is capable to obtain the quantitative and clear surface friction line map and oil film velocity field. Therefore, its advantages are obvious comparing with the tradition methods, and its engineering application value is high.

Key words: global surface friction, friction line map, movement decoupling, discrete matching, fluorescent oil film