Enhanced particle-tracking velocimetry (EPTV) with a combined two-component pair-matching algorithm

Abstract

The main goal of this paper is to present, validate and demonstrate recent improvements to the original particle identification and tracking technique PTV (particle-tracking velocimetry), named EPTV (enhanced particle-tracking velocimetry). In order to improve the performance of the image-processing tools used in EPTV by means of particle-size-based tracking, a new combined two-component pair-matching algorithm has been developed, using both particle-size-related data and data for displacements of possible particles moving similarly within the neighbouring environment. Significant technique improvements have been successfully validated on synthetic images and demonstrated on real-world images. Results show the algorithm's capability to provide high spatial resolution over a wide velocity dynamic range in turbulent flow measurements.

Relaxation algorithm-based PTV with dual calculation method and its application in addressing particle saltation

“…The test is conducted using synthetic particle images of PIV Standard Image project (Okamoto et al 2000a, b), which has been extensively used to verify PTV/PIV algorithms (Brevis et al 2011;Cardwell et al 2011;Mikheev and Zubtsov 2008;Ohmi and Li 2000;Ohmi and Panday 2009;Ohmi et al 2010). Three sets of 2D image series, numbered as 01, 04 and 23, are used here.…”

Iterative particle matching for three-dimensional particle-tracking velocimetry

“…First, an overdetermined third-order polynomial spatial interpolation method is used for the dewarping step resulting in a smooth dewarping displacement distribution, which dampens the effect of possibly invalid spurious matches. Second, an iterative unstructured median test is applied after each matching pass similar to Mikheev and Zubtsov (2008). The potential loss of some valid displacements is favorable to a potential spurious prediction, which could prevent the convergence completely, since even a coarse and imprecise prediction and dewarping step improves the tracking parameter by reducing the remaining displacement.…”

“…The relaxation approach achieved a significantly better matching performance. Mikheev and Zubtsov (2008) proposed to use the particle intensity to identify similar particles across exposures to have more information for the correct identification of particle correspondences. While this method can be applied in twodimensional measurements, its suitability for volumetric flows has yet to be shown.…”

Experimental determination of droplet collision rates in turbulence

“…The particle segmentation relies on a modification of the dynamic threshold algorithm proposed by Mikheev and Zubtsov [46]. Two thresholds are required [47]: firstly, an absolute intensity threshold for peak detection and secondly, a dynamic threshold for pixel-particle allocation.…”