Dynamic masking of PIV images using the Radon transform in free surface flows

Abstract

Particle image velocimetry (PIV) processing of free surface flow images often requires the use of digital masks to overcome the problems caused by the interface. In cases where a large number of particle images are collected it is essential that the time-varying boundary between the two phases can be tracked automatically to produce the binary masks. The Radon transform-based technique presented in this paper allows the automatic detection of the air-water interface in a stream of particle images acquired from a single camera. It is applied to time-resolved PIV measurements in the liquid phase of a stratified multiphase flow in a circular pipe. Accuracy estimations are provided using synthetic and real wave profiles. An extension to the more complex case of an overturning wave is also discussed.

Measurement of the velocity field in parametrically excited solitary waves

“…Since image sequences included a moving boundary, we required an automatic algorithm for boundary detection. We achieved this by using a Radon-transformbased method from Sanchis & Jensen (2011) on averaged same-phase samples. Likewise, background and illumination issues were corrected using statistics-based images.…”

Experiments on generation of surface waves by an underwater moving bottom

“…In the images, the particles form an identifiable line, which can be used for detection (figure 2). The free-surface vertical deformation η(r, t) was then obtained by applying a Radon transform algorithm on the images [39]. This algorithm integrates the intensity along all the possible straight lines contained in a sub-window and finds the maximal value.…”

A new dynamic masking technique for time resolved PIV analysis

“…However, when the position and/or shape of the solid objects varies in time in an unknown way, a masking technique able to dynamically track the object is needed. The dynamic masking (DM) approach for PIV analysis is currently receiving a considerable amount of attention (Sanchis and Jensen 2011;Masullo and Theunissen 2017;Anders et al 2019), thanks to the diffusion of time resolved PIV systems which, in turn, benefit from the increased availability of high speed cameras. The major advancements in DM techniques come from the field of micro-PIV (Lindken et al 2009) applied to microfluidics, where the investigation of both the flow fields around micro-and nanoswimmers (Ergin et al 2015) and multiphase flows (Brücker 2000;Khalitov and Longmire 2002) requires accurate and flexible algorithms.…”