Real-time image processing for particle tracking velocimetry

We present a novel high-speed particle tracking velocimetry (PTV) experimental system. Its novelty is due to the FPGA-based, real-time image processing “on camera”. Instead of an image, the camera transfers to the computer using a network card, only the relevant information of the identified flow tracers. Therefore, the system is ideal for the remote particle tracking systems in research and industrial applications, while the camera can be controlled and data can be transferred over any high-bandwidth network. We present the hardware and the open source software aspects of the PTV experiments. The tracking results of the new experimental system has been compared to the flow visualization and particle image velocimetry measurements. The canonical flow in the central cross section of a a cubic cavity (1:1:1 aspect ratio) in our lid-driven cavity apparatus is used for validation purposes. The downstream secondary eddy (DSE) is the sensitive portion of this flow and its size was measured with increasing Reynolds number (via increasing belt velocity). The size of DSE estimated from the flow visualization, PIV and compressed PTV is shown to agree within the experimental uncertainty of the methods applied.     

New tracking algorithm for particle image velocimetry

The cross correlation tracking technique is widely used to analyze image data, in Particle Image Velocimetry (PIV). The technique assumes that the fluid motion, within small regions of the flow field, is parallel over short time intervals. However, actual flow fields may have some distorted motion, such as rotation, shear and expansion. Therefore, if the distortion of the flow field is not negligible, the fluid motion can not be tracked well using the cross correlation technique. In this study, a new algorithm for particle tracking, called the Spring Model technique, has been proposed. The algorithm can be applied to flow fields which exhibit characteristics such as rotation, shear and expansion.The algorithm is based on pattern matching of particle clusters between the first and second image. A particle cluster is composed of particles which are assumed to be connected by invisible elastic springs. Depending on the deformation of the cluster pattern (i.e., the particle positions), the invisible springs have some forces. The smallest force pattern in the second image is the most probable pattern match to the correspondent original pattern in the first image. Therefore, by finding the best matches, particle movements can be tracked between the two images. Three-dimensional flow fields can also be reconstructed with this technique.The effectiveness of the Spring Model technique was verified with synthetic data from both a two-dimensional flow and three-dimensional flow. It showed a high degree of accuracy, even for the three-dimensional calculation. The experimental data from a vortex flow field in a cylinder wake was also measured by the Spring model technique.     

High-precision sub-pixel interpolation in particle image velocimetry image processing

The task of image interpolation and re-sampling for particle image velocimetry (PIV) is investigated, which is used for window shifting with sub-pixel accuracy and image or window deformation. A new interpolation scheme based on a Gaussian filter is introduced and compared with commonly used and widely accepted interpolation techniques in terms of the achievable root mean square deviation of the displacement estimates.     

2-Dimensional particle tracking velocimetry (PTV): Technique and image processing algorithms

An automated particle track velocimeter (PTV) was constructed to determine the fluid velocity field in a transparent test section (an engine throttle body assembly is used as an example) by analyzing images of scattered light from hollow nylon particles as they move with the flow. The light from individual particles was imaged on an SIT vidicon at video rates, digitized on an FG100CD frame buffer board and analyzed automatically in an IBMPC/AT. The essence of this technique is a novel processing algorithm which converts particle track segments to flow velocity vectors at the rate of 2–3 images/minute with this hardware. An application of the technique to a gas phase flow ( 6.5 m/s) through this throttle body assembly is presented.     

Tomographic particle tracking velocimetry using telecentric imaging

The goal of this article is to discuss 3D Particle Tracking Velocimetry (PTV) in a tomographic reconstructed voxel space with at least doubling the spatial resolution compared to classical 3D PTV. For this purpose, a new tomographic reconstruction technique based on telecentric imaging in combination with the epipolar geometry is presented. The method overcomes the need for memory intensive weighting matrices or cost intensive iterations, which are necessary in iterative algebraic reconstruction techniques. A characteristic of tomographic reconstruction is the reconstruction of ghost particles. As the aim of PTV is the reconstruction of true particle paths, this article focuses on the removal of ghost particles and ghost trajectories. The method is validated via a synthetic turbulent flow field and via the benchmark experiment of a vortex ring.     

Two-dimensional Gaussian regression for sub-pixel displacement estimation in particle image velocimetry or particle position estimation in particle tracking velocimetry

An explicit solution of two-dimensional Gaussian regression for the estimation of particle displacement from the correlation function in particle image velocimetry (PIV) or particle position from the images in particle tracking velocimetry (PTV) with sub-pixel accuracy is introduced. The accuracy and the ability of the methods to avoid pixel locking due to non-axially orientated, elliptically shaped particle images or correlation peaks are investigated using simulated and experimentally obtained images.     

Three-dimensional particle tracking method using FPGA-based real-time image processing and four-view image splitter

We present a cost-effective solution of the three-dimensional particle tracking velocimetry (3D-PTV) system based on the real-time image processing method (Kreizer et al. Exp Fluids 48:105–110, 2010) and a four-view image splitter. The image processing algorithm, based on the intensity threshold and intensity gradients estimated using the fixed-size Sobel kernel, is implemented on the field-programmable gate array integrated into the camera electronics. It enables extracting positions of tracked objects, such as tracers or large particles, in real time. The second major component of this system is a four-view split-screen device that provides four views of the observation volume from different angles. An open-source ray-tracing software package allows for a customized optical setup for the given experimental settings of working distances and camera parameters. The specific design enables tracking in larger observation volumes when compared to the designs published up to date. The present cost-effective solution is complemented with open-source particle tracking software that receives raw data acquired by the real-time image processing system and returns trajectories of the identified particles. The combination of these components simplifies the 3D-PTV technique by reducing the size and increasing recording speed and storage capabilities. The system is capable to track a multitude of particles at high speed and stream the data over the computer network. The system can provide a solution for the remotely controlled tracking experiments, such as in microgravity, underwater or in applications with harsh experimental conditions.     

Investigation of particle-laden turbulent pipe flow at high-Reynolds-number using particle image/tracking velocimetry (PIV/PTV)

An experimental investigation of a high Reynolds number flow (Re = 320 000) of a dilute liquid-solid mixture (<1% by volume) was conducted. The turbulent motion of both the liquid phase (water) and particles (0.5, 1, and 2 mm glass beads) was evaluated in an upward pipe flow using a particle image/tracking velocimetry (PIV/PTV) technique. Results show that the Eulerian mean axial velocity of the glass beads is lower than that of the liquid phase in the central region but higher in the near-wall region. Moreover, the presence of the coarse particles has a negligible effect on the turbulence intensity of the liquid phase. Particles show higher streamwise and radial fluctuations than the liquid-phase at the tested conditions. The profiles of particle concentration across the pipe radius show almost constant concentration in the core of the pipe with a decrease towards the near wall region for 0.5 and 1 mm particles. For the 2 mm particles, a nearly linear concentration gradient from centre to the pipe wall is observed. The results presented here provide new information concerning the effect of a dispersed particulate phase on the turbulence modulation of the liquid carrier phase, especially at high Reynolds numbers. The present study also demonstrates how correlations developed to determine if particles cause turbulence attenuation/augmentation are not applicable for solid-liquid flows at high Reynolds numbers. Finally, the importance of particle-fluid slip velocity on fluid phase turbulence modulation is illustrated.     

A new artificial neural network tracking technique for particle image velocimetry

 The analysis of Particle Image Velocimetry (PIV) data requires effective algorithms to track efficiently the particles suspended in the fluid flow. The artificial neural network algorithm method described here presents a new approach to solve this problem. Contrary to the classic cross correlation method, this new method does not require a large number of particles per frame, it can handle flows with large velocity gradients, and is suited for tracking images with multiple exposures as well as tracking through consecutive images. The algorithm was tested on synthetic and available experimental data to provide a thorough performance analysis. Received: 28 May 1996/Accepted: 25 December 1996     

Stereoscopic particle image velocimetry

Stereoscopic particle image velocimetry (PIV) employs two cameras to record simultaneous but distinct off-axis views of the same region of interest (illuminated plane within a flow seeded with tracer particles). Sufficient information is contained in the two views to extract the out-of-plane motion of particles, and also to eliminate perspective error which can contaminate the in-plane measurement. This review discusses the principle of stereoscopic PIV, the different stereoscopic configurations that have been used, the relative error in the out-of-plane to the in-plane measurement, and the relative merits of calibration-based methods for reconstructing the three-dimensional displacement vector in comparison to geometric reconstruction. It appears that the current trend amongst practitioners of stereoscopic PIV is to use digital cameras to record the two views in the angular displacement configuration while incorporating the Scheimpflug condition. The use of calibration methods has also gained prominence over geometric reconstruction. Received: 15 April 1999/Accepted: 1 February 2000     

Tomographic particle image velocimetry

This paper describes the principles of a novel 3D PIV system based on the illumination, recording and reconstruction of tracer particles within a 3D measurement volume. The technique makes use of several simultaneous views of the illuminated particles and their 3D reconstruction as a light intensity distribution by means of optical tomography. The technique is therefore referred to as tomographic particle image velocimetry (tomographic-PIV). The reconstruction is performed with the MART algorithm, yielding a 3D array of light intensity discretized over voxels. The reconstructed tomogram pair is then analyzed by means of 3D cross-correlation with an iterative multigrid volume deformation technique, returning the three-component velocity vector distribution over the measurement volume. The principles and details of the tomographic algorithm are discussed and a parametric study is carried out by means of a computer-simulated tomographic-PIV procedure. The study focuses on the accuracy of the light intensity field reconstruction process. The simulation also identifies the most important parameters governing the experimental method and the tomographic algorithm parameters, showing their effect on the reconstruction accuracy. A computer simulated experiment of a 3D particle motion field describing a vortex ring demonstrates the capability and potential of the proposed system with four cameras. The capability of the technique in real experimental conditions is assessed with the measurement of the turbulent flow in the near wake of a circular cylinder at Reynolds 2,700.     

Digital particle image velocimetry

Digital particle image velocimetry (DPIV) is the digital counterpart of conventional laser speckle velocitmetry (LSV) and particle image velocimetry (PIV) techniques. In this novel, two-dimensional technique, digitally recorded video images are analyzed computationally, removing both the photographic and opto-mechanical processing steps inherent to PIV and LSV. The directional ambiguity generally associated with PIV and LSV is resolved by implementing local spatial cross-correlations between two sequential single-exposed particle images. The images are recorded at video rate (30 Hz or slower) which currently limits the application of the technique to low speed flows until digital, high resolution video systems with higher framing rates become more economically feasible. Sequential imaging makes it possible to study unsteady phenomena like the temporal evolution of a vortex ring described in this paper. The spatial velocity measurements are compared with data obtained by direct measurement of the separation of individual particle pairs. Recovered velocity data are used to compute the spatial and temporal vorticity distribution and the circulation of the vortex ring.     

Nonlinear temporal filtering of time-resolved digital particle image velocimetry data

Nonlinear filtering methods have been developed to identify and replace outlying data points in velocity time series obtained with time-resolved digital particle image velocimetry (PIV) of the flow around a surface-mounted cube at a Reynolds number of 20,000. Nuances associated with the spectral computation of the cross-correlation are highlighted, including the requirement of zero-padding an image interrogation area to eliminate the circular components of the cross-correlation. Three nonlinear filtering methods for the replacement of outliers are applied to the velocity time series sampled at 1,000 Hz: a median filter, a decision-based Hampel filter, and a PIV-specific Hampel filter. The particular benefit of the PIV-specific Hampel filter is that it allows the retention of actual measured data, sometimes derived from alternate peaks in the cross-correlation function, while still providing for the removal of outliers when a consistent, nonoutlying measurement is not available.     

Efficient detection of spurious vectors in particle image velocimetry data

A statistical model is introduced that describes the occurence of spurious vectors in PIV data. This model is used to investigate the performance of three different post-interrogation procedures: the global-mean, the local-mean and the local-median test. The model is also used to optimize the performance of these procedures. Predicted performances agree very well with those obtained from an artificially generated PIV record. It is demonstrated that the detectability as the conventional measure for the reliability of a measured displacement vector is very inefficient, compared to the three tests described here. The local-median test has the highest efficiency.     

Stereoscopic micro particle image velocimetry

A stereoscopic micro-PIV (stereo-μPIV) system for the simultaneous measurement of all three components of the velocity vector in a measurement plane (2D–3C) in a closed microchannel has been developed and first test measurements were performed on the 3D laminar flow in a T-shaped micromixer. Stereomicroscopy is used to capture PIV images of the flow in a microchannel from two different angles. Stereoscopic viewing is achieved by the use of a large diameter stereo objective lens with two off-axis beam paths. Additional floating lenses in the beam paths in the microscope body allow a magnification up to 23×. The stereo-PIV images are captured simultaneously by two CCD cameras. Due to the very small confinement, a standard calibration procedure for the stereoscopic imaging by means of a calibration target is not feasible, and therefore stereo-μPIV measurements in closed microchannels require a calibration based on the self-calibration of the tracer particle images. In order to include the effects of different refractive indices (of the fluid in the microchannel, the entrance window and the surrounding air) a three-media-model is included in the triangulation procedure of the self-calibration. Test measurement in both an aligned and a tilted channel serve as an accuracy assessment of the proposed method. This shows that the stereo-μPIV results have an RMS error of less than 10% of the expected value of the in-plane velocity component. First measurements in the mixing region of a T-shaped micromixer at Re = 120 show that 3D flow in a microchannel with dimensions of 800 × 200 μm² can be measured with a spatial resolution of 44 × 44 × 15 μm³. The stationary flow in the 200 μm deep channel was scanned in multiple planes at 22 μm separation, providing a full 3D measurement of the averaged velocity distribution in the mixing region of the T-mixer. A limitation is that this approach requires a stereo-objective that typically has a low NA (0.14–0.28) and large depth-of-focus as opposed to high NA lenses (up to 0.95 without immersion) for standard μPIV.     

High image-density particle image velocimetry using laser scanning techniques

Laser scanning, corresponding to time-dependent deflections of laser beam across a field of interest, can provide relatively high illumination intensity of small particles, thereby allowing implementation of high image-density particle image velocimetry (PIV). Scanning techniques employing a rotating (multi-faceted) mirror, an oscillating mirror, and an acousto-optic deflector are addressed. Issues of illumination intensity and exposure, rate of scan of the laser beam, and retrace time of the scanning beam are assessed. Representative classes of unsteady separated flows investigated with laser-scanning PIV are described.     

Large-scale tomographic particle image velocimetry using helium-filled soap bubbles

To measure large-scale flow structures in air, a tomographic particle image velocimetry (tomographic PIV) system for measurement volumes of the order of one cubic metre is developed, which employs helium-filled soap bubbles (HFSBs) as tracer particles. The technique has several specific characteristics compared to most conventional tomographic PIV systems, which are usually applied to small measurement volumes. One of them is spot lights on the HFSB tracers, which slightly change their position, when the direction of observation is altered. Further issues are the large particle to voxel ratio and the short focal length of the used camera lenses, which result in a noticeable variation of the magnification factor in volume depth direction. Taking the specific characteristics of the HFSBs into account, the feasibility of our large-scale tomographic PIV system is demonstrated by showing that the calibration errors can be reduced down to 0.1 pixels as required. Further, an accurate and fast implementation of the multiplicative algebraic reconstruction technique, which calculates the weighting coefficients when needed instead of storing them, is discussed. The tomographic PIV system is applied to measure forced convection in a convection cell at a Reynolds number of 530 based on the inlet channel height and the mean inlet velocity. The size of the measurement volume and the interrogation volumes amount to 750 mm × 450 mm × 165 mm and 48 mm × 48 mm × 24 mm, respectively. Validation of the tomographic PIV technique employing HFSBs is further provided by comparing profiles of the mean velocity and of the root mean square velocity fluctuations to respective planar PIV data.     

Hybrid stereoscopic particle image velocimetry

 The technical aspects of a photographic stereo camera for three-dimensional particle image velocimetry are described herein. The hybrid concept of the camera combines advantages of the angular displacement and the translation method. The camera uses two CCD sensors in order to adjust the lens distances and angles to meet the Scheimpflug criterion and two coupled rotating mirrors for image shifting. An application to a jet flow with an exit velocity of 33 m/s demonstrates the succesfull optimization of the recording process. Received: 27 September 1996/Accepted: 6 March 1997     

Estimation of particle dynamics in 2-D fluidized beds using particle tracking velocimetry

The experimental characterization of particle dynamics in fluidized beds is of great importance in fostering an understanding of solid phase motion and its effect on particle properties in granulation processes. Commonly used techniques such as particle image velocimetry rely on the cross-correlation of illumination intensity and averaging procedures. It is not possible to obtain single particle velocities with such techniques. Moreover, the estimated velocities may not accurately represent the local particle velocities in regions with high velocity gradients. Consequently, there is a need for devices and methods that are capable of acquiring individual particle velocities. This paper describes how particle tracking velocimetry can be adapted to dense particulate flows. The approach presented in this paper couples high-speed imaging with an innovative segmentation algorithm for particle detection, and employs the Voronoi method to solve the assignment problem usually encountered in densely seeded flows. Lagrangian particle tracks are obtained as primary information, and these serve as the basis for calculating sophisticated quantities such as the solid-phase flow field, granular temperature, and solid volume fraction. We show that the consistency of individual trajectories is sufficient to recognize collision events.