Understanding the complex fluid-particle interactions in shear-induced erosion and scour requires measurement of the interstitial fluid motion in both stationary and moving particle beds. These measurements are confounded by the differing material properties of the fluids and particles used in experiments. In the present investigation, full-field fluid-velocity measurements for flow over a bed of 3-mm-diameter monodisperse glass beads were obtained using digital particle image velocimetry (DPIV) in an aqueous solution of sodium iodide to match the refractive index of the particles. Fluorescent seeding particles were used in the fluid together with optical filters and image processing to provide both time-averaged and instantaneous measurements of the fluid velocity for both stationary and moving particle beds at channel Reynolds numbers (Re) in the range 4900–9000. Recent investigations have used refractive index matched fluids to obtain flow field measurements in the particle bed, but have been limited to time-averaged measurements, stationary particle beds, and/or low Re. The results indicate that particle motion was initiated at a critical Shields number of approximately 0.017, but some hysteresis in the Shields number was observed once motion was initiated. Velocity profiles were compared with power law and log-law models, showing reasonable agreement in certain flow regimes, but the particle-layer shear stress predicted by a log-law fit was found to be 10–30 times the actual shear stress measured from DPIV measurements, suggesting measurements of particle shear in channel flow based on a log-law model may not be sufficiently accurate. A correction to the log-law model utilized in porous wall flows was applied, and the adjusted value of von Karman constant increased with the shear Reynolds number for the cases without particle motion, but no clear trend was observed for the moving particle cases.