DPIV (Digital Particle-Imaging Velocimetry) has been attached importance to and applied wisely interiorly and overseas. But at present

its most difficult problem is the precision. Because the tracer particle images of DPIV are sampled by CCD and image digitizer

the noise (mainly the noise of particle image size

the size of interrogation window

local velocity gradients

the number of particles within the sampling window and quantization effects) that is imported inevitably in the process of experiment falls the precision of DPIV experiment. This paper presents the image (simulative and practical image) noise removing in DPIV based on wavelet transform whose characteristic is the multiresolution. This method is compared with image noise removing by Wiener and median filter. The result shows that image noise removing in DPIV based on wavelet transform improves the precision of DPIV experiment

and it is the most precise to rebuild the velocity field based on cross-correlation after image noise removing in DPIV using wavelet transform.