We present a new Image Velocimetry (IV) hybrid that estimates vector fields from images with widely different visualization pattern sizes such as those encountered in riverine bedform migration. The IV approach is obtained by complementing the Cross-correlation Method (CCM) with algorithms of Optical Flow Methods (OFM). The OFM procedures are first applied to automatically determine the optimal Search Windows (SWs) over the whole imaged area. Subsequently, the CCM utilized the established SWs to locally resolve velocity fields associated with the bedform movement. The new approach, labeled herein as High-Gradient Pattern IV (HGPIV), combines the advantages of both parent techniques to improve the accuracy and spatial resolution of the resultant global velocity field and significantly reduces the computational time. The HGPIV validation consists of comparing its results with those obtained with the CCM approaches applied for estimating the velocity field associated with bedform migration in a large river.
