TY - JOUR
T1 - PIV algorithms for open-channel turbulence research
T2 - accuracy, resolution and limitations
AU - Cameron, Stuart M.
PY - 2011/12
Y1 - 2011/12
N2 - The implementation of PIV for experimental studies in open-channel flows can be challenging due to the presence of strong velocity gradients and the inclusion of solid interfaces in the captured images. Understanding the performance and limitations of the PIV method under these conditions is critical for optimising experimental parameters and robust interpretation of data. The optimum algorithm for extracting velocity fields from PIV images is the subject of ongoing revision with the goal of maximising resolution and minimising errors, and recent advances in this regard may be particularly beneficial for open-channel turbulence research. Key steps in the iterative discrete shift (IDS) and image deformation method (IDM) algorithms are detailed, and the fundamental differences between direct cross correlation and FFT-based correlation methods are explained. It is also shown how the resolution of an algorithm can be determined from its modulation transfer function (MTF), and how the MTF can be manipulated with the selection of intensity weighting windows. The random error levels for selected algorithms are demonstrated under different image and flow field conditions, including the near boundary region, using simulated PIV images.
AB - The implementation of PIV for experimental studies in open-channel flows can be challenging due to the presence of strong velocity gradients and the inclusion of solid interfaces in the captured images. Understanding the performance and limitations of the PIV method under these conditions is critical for optimising experimental parameters and robust interpretation of data. The optimum algorithm for extracting velocity fields from PIV images is the subject of ongoing revision with the goal of maximising resolution and minimising errors, and recent advances in this regard may be particularly beneficial for open-channel turbulence research. Key steps in the iterative discrete shift (IDS) and image deformation method (IDM) algorithms are detailed, and the fundamental differences between direct cross correlation and FFT-based correlation methods are explained. It is also shown how the resolution of an algorithm can be determined from its modulation transfer function (MTF), and how the MTF can be manipulated with the selection of intensity weighting windows. The random error levels for selected algorithms are demonstrated under different image and flow field conditions, including the near boundary region, using simulated PIV images.
KW - particle image velocimetry
KW - resolution
KW - error
KW - accuracy
UR - http://www.scopus.com/inward/record.url?scp=80455156055&partnerID=8YFLogxK
U2 - 10.1016/j.jher.2010.12.006
DO - 10.1016/j.jher.2010.12.006
M3 - Article
AN - SCOPUS:80455156055
SN - 1570-6443
VL - 5
SP - 247
EP - 262
JO - Journal of Hydro-Environment Research
JF - Journal of Hydro-Environment Research
IS - 4
ER -