In order to eliminate directional ambiguity and to reduce standard uncertainty in velocity measurement, a novel spatial filtering velocimeter based on a spatiotemporal filter is established. The spatiotemporal filter is structured by a high speed complementary metal oxide semiconductors (CMOS) linear image sensor. The spatial filtering characteristic of the filter is analyzed theoretically. The power spectrum of the output signal is obtained by fast Fourier transform (FFT), and is corrected by a frequency spectrum correction algorithm, named energy centrobaric correction. Therefore, the standard uncertainty of frequency measurement is reduced. This novel velocimeter is used to measure the moving velocity of a conveyor belt. The experiment verifies the feasibility of the spatiotemporal filter for directional discrimination. Furthermore, the relative standard uncertainty of velocity measurement is analyzed experimentally. Experimental results of the presented spatial filtering velocimeter using spatiotemporal filter show that the relative standard uncertainty of velocity measurement is much reduced compared with that of frequency measurement. The reduced relative standard uncertainty of velocity measurement is below 0.24% , which can be further reduced by adjusting the parameters of the spatiotemporal filter.