The accurate evaluation of grinding wheel surface topography, which is necessary for the investigation of the grinding principle, optimism, modeling, and simulation of a grinding process, significantly depends on the accurate recognition of abrasive grains from the measured wheel surface. A detailed analysis of the grain size distribution characteristics and grain profile wavelength of the fine diamond grinding wheel used for ultra-precision grinding is presented. The requirements of the spatial sampling interval and sampling area for instruments to measure the surface topography of a diamond grinding wheel are discussed. To recognize diamond grains, digital filtering is used to eliminate the high frequency disturbance from the measured 3D digital surface of the grinding wheel, the geometric features of diamond grains are then extracted from the filtered 3D digital surface, and a method based on the grain profile frequency characteristics, diamond grain curvature, and distance between two adjacent diamond grains is proposed. A 3D surface profiler based on scanning white light interferometry is used to measure the 3D surface topography of a #3000 mesh resin bonded diamond grinding wheel, and the diamond grains are then recognized from the 3D digital surface. The experimental result shows that the proposed method is reasonable and effective.