When large fields of view are used with particle image velocimetry (PIV) in thestudy of complex fluid flows, extraneous effects linked to velocity gradients andnon-uniformities in both image illumination and particle number density becomemore prevalent. These factors, coupled with the limiting requirement that largeareas of interest (AOIs) must be employed to measure the full range of velocity,cause degradation of correlation results (i.e. broadening and/or splintering of thecross-correlation peaks). Advanced iterative and hierarchical PIV strategies provideimproved results but these can break down in complex flows where velocity gradients aresignificant and particle dispersion does not remain uniformly random. One reasonfor this breakdown is that local displacement vectors obtained using the crosscorrelation method are not necessarily representative of the fluid motion wherethese vectors are typically anchored (namely, the geometric centre of the AOI).To address this issue a simple but effective technique is presented that enablesindividual displacement vectors to be anchored within an AOI at a location(s)where the actual fluid motion is more consistent with the measured displacement.The method involves a straightforward approach to extract the intensity featuresfrom within each AOI that most influence the calculation of the cross-correlationplane. To demonstrate the utility of the methodology, bounds of uncertainty areapproximated, and results obtained from the analysis of high gradient synthetic flowfields are compared against results obtained using the conventional PIV approach.