Abstract: The particle image velocimetry (PIV) is an effective and non-intrusive technique to measure the planar distribution of velocity in the fluid based on the cross-correlation of flow images.In general,cross-correlation analysis in the PIV can be implemented quickly by the fast Fourier transform,and the efficiency can be improved by considering the property of real Fourier transform.Furthermore,the higher efficiency can be obtained by using the Hartley transform of separable kernels.Three algorithms are specified in detail in the paper,and the comparison of the three algorithms is conducted in the theory and computation time in practice,which can show the advantages of the Hartley transform.