Digital Filters for Reducing Background Noise in Micro PIV Measurements

In the present work digital filters are used to reduce the measurement uncertainty of the microscopic particle image velocimetry (µPIV) technique. Analysis shows that single pixel random noise and low frequency background noise in the µPIV recordings are much stronger than in the usual macroscopic PIV recordings. The effects of several different low- and high-pass filters on the evaluation accuracy of a correlation-based interrogation algorithm with continuous window shift and central difference image correction are investigated. Simulations show that the 3×3-pixel rectangular smooth filter can reduce effectively the evaluation error resulting from single pixel random noise. The unsharp mask is a powerful tool to reduce the evaluation error caused by low frequency background noise. A µPIV image filter that combines the capabilities of both the smooth filter and the unsharp mask is constructed to enable high accuracy evaluation of µPIV recordings. Fig. 1: µPIV evaluation sample pair (64×64 pixels) mn Φ(m,n) ( a