Yes, it is low. But I am working on the open channel flow with a mean velocity of 42 cm/s. Further, I have taken a large number of images to compensate for the low frequency.
You cannot compensate using a large number of images, the sampling frequency of 8Hz will cut away the most relevant part of your wavenumber content...
If I understand, you get 0.42m/s of average velocity and you have water, right? What about the characteristic lenght dimension to estimate the Re number? Have you estimated the Kolmogorov frequency to evaluate where your Nyquist cut-off lie?