The detrimental effect of increasing the number of cameras on self-calibration for tomographic PIV

Abstract

It is commonly assumed that adding cameras is beneficial for the accuracy of multiple-camera systems. This tautological assertion is most certainly true for tomographic particle image velocimetry and 3D particle-tracking velocimetry systems, where the search area on each image for the particle corresponding to a 3D trial position is small (typically less than the particle image diameter). On the other hand, when it comes to larger search areas (due to, for example, calibration uncertainties), quite surprisingly, increasing the number of cameras might have a detrimental effect. Under some conditions, this is the case for the volumetric self-calibration technique (Wieneke 2008 Exp. Fluids 45 549–56), in which the residual calibration error and misalignments of the cameras are corrected by statistically searching matching particles over a search area larger than the expected maximum calibration error. In this work, the loss of signal in the self-calibration for systems with three or more cameras is discussed. Two readily implementable solutions are provided to reduce this source of error. The algorithms are tested on synthetic and real images.