Large-scale Particle Image Velocimetry Research Articles

A method of applying deep learning based optical flow algorithm to river flow discharge measurement

River flow discharge monitoring is one of the critical tasks performed at hydrological stations. The large-scale particle image velocimetry (LSPIV) method widely used in hydrological stations is often limited by a lack of floating objects and has a high computational complexity. The space-time image velocimetry method is susceptible to noise interference and requires high stability of the flow over time. This paper proposes a flow measurement method based on the recurrent all-pairs field transforms for optical flow (RAFT) algorithm. The 4D correlation volume pyramid in the RAFT network structure can effectively handle changing and complex flow conditions. The convolutional block attention module is introduced into the optical flow update module after the 4D correlation volume pyramid, enhancing the ability to capture complex flow surface information. Additionally, feature extraction adds deformable convolution to expand the receptive field of the flow image, which has better adaptability in non-rigid motion. To validate the effectiveness of the new method (RAFT-D-C), this paper conducts comparative experiments with both existing and new methods. The experimental results show that RAFT-D-C has relative errors of 2.13% and 4.41% for the average flow velocity of two rivers and relative errors of 2.19% and 3.05% for the total discharge, respectively. RAFT-D-C provides improved accuracy compared to other methods and requires less computational run time than the frequently used LSPIV method.

Natural surface floaters in image-based river surface velocimetry: Insights from a case study

This study focuses on utilizing image techniques for river velocity measurement, with a specific emphasis on natural surface floating patterns. Employing a multi-camera system, we conducted 3D measurements on river surfaces, including surface velocity and water surface reconstruction. A pattern-based tracking approach has been adopted to improve the performance of image measurements on different types of natural floating tracers. The study employs the following approaches: 3D Lagrangian Pattern Tracking Velocimetry (3D-LPTV), 2D Lagrangian Pattern Velocimetry (2D- LPTV), and Large-scale Particle Image Velocimetry (LSPIV), for surface velocity estimation. The outcomes revealed that all three approaches yielded consistent results in terms of averaged velocity. However, the LSPIV method produced about two times higher uncertainty in measured velocities compared to the other methods. A strategy to assess the quality of river surface patterns in velocity estimation is presented. Specifically, the sum of squared interrogation area intensity gradient (SSIAIG) was found to be strongly correlated with measurement uncertainty. Additionally, a term related to the peak sidelobe ratio (PSR) of the cross-correlation map was found as an effective constraint, ensuring the image-tracking process achieves high reliability. The precision of measurements increases corresponding to the increase of image intensity gradient and PSR.

Surface Current Measurement Using Large Scale Particle Image Velocimetry (LSPIV) in Ender Estuary, Cirebon

The Large-Scale Particle Image Velocimetry (LSPIV) method is an image- based technique that uses a non-intrusive approach or avoids direct physical contact with the water being studied to measure the velocity of water flow. The present study aims to identify surface velocity using LSPIV captured by Unmanned Aerial Vehicle (UAV) in Ender estuary, Cirebon. Field observation for surface velocity was conducted at three locations during flood and ebb tides. The surface current results of LSPIV compared with in situ velocity measurement by a current meter. The results of the LSPIV method are strongly influenced by the data acquisition process such as lighting, interrogation area, and video recording angle. Based on comparing the surface flow between results of LSPIV and observation data, the coefficient correlation (R) is 0.84 for velocity magnitude and R = 0.92 for flow direction. The Root Mean Square Error (RMSE) of 0.024 m/s for velocity magnitude and 4.81 degree for flow direction, and the Mean Absolute Error (MAE) range from 0.002 to 0.059 m/s (1 to 12 degree) for the velocity magnitude and the flow direction, respectively. Furthermore, the Mean Absolute Percentage Error (MAPE) value of 10.41% (8.84%) for magnitude and direction, respectively. Therefore, LSPIV was applied successfully to obtain surface velocity measurements.

Quantifying and Reducing the Operator Effect in LSPIV Discharge Measurements

AbstractOperator choices, both in acquiring the video and data and in processing them, can be a prominent source of error in image‐based velocimetry methods applied to river discharge measurements. The Large Scale Particle Image Velocimetry (LSPIV) is known to be sensitive to the parameters and computation choices set by the user, but no systematic comparisons with discharge references or intercomparisons have been conducted yet to evaluate this operator effect in LSPIV. In this paper, an analysis of a video gauging intercomparison, the Video Globe Challenge 2020, is proposed to evaluate such operator effect. The analysis is based on the gauging reports of the 15 to 23 participants using the Fudaa‐LSPIV software and intents to identify the most sensitive parameters for the eight videos. The analysis highlighted the significant impact of the time interval, the grid points and the filters on the LSPIV discharge measurements. These parameters are often inter‐dependent and should be correctly set together to strongly reduce the discharge errors. Based on the results, several automated tools were proposed to reduce the operator effect. These tools consist of several parameter assistants to automatically set the orthorectification resolution, the grid and the time interval, and of a sequence of systematic and automatic filters to ensure reliable velocity measurements used for discharge estimation. The application of the assisted LSPIV workflow using the proposed tools leads to significant improvements of the discharge measurements with strong reductions of the inter‐participant variability. On the eight videos, the mean interquartile range of the discharge errors is reduced from 17% to 5% and the mean discharge bias is reduced from −9% to 1% with the assisted LSPIV workflow. The remaining inter‐participant variability is mainly due to the user‐defined surface velocity coefficient α.

On the performance of streamflow gauging using CCTV-integrated LSPIV in diverse hydro-environmental conditions.

Addressing the critical need for precise streamflow measurements in hydro-environmental research, this study evaluates large-scale particle image velocimetry (LSPIV) using cost-effective closed-circuit television (CCTV) cameras, providing a detailed sensitivity analysis in both laboratory and real-world canal settings. In laboratory conditions, a 45° camera angle notably enhanced performance, exhibiting a 12% decrease in MAE and a remarkable 40% reduction in RMSE compared to the performance of orthographic form. Tracer particles further enhanced LSPIV accuracy, decreasing both mean absolute error (MAE) and root mean square error (RMSE) by around 0.05 m/s. Optimal velocity coefficients for the lab ranged between 0.85 and 0.90. Nighttime measurements, using projection-based illumination, showed a minor 3% MAE variation and 0.02 RMSE difference versus daytime. In field experiments, a 45° upstream CCTV camera configuration notably improved LSPIV accuracy, achieving a 3% MAE and 0.055 m/s RMSE. For best results across different turbidity levels, we recommend a velocity coefficient range of 0.84 to 0.88. This study highlights the robustness and cost-efficiency of LSPIV as a transformative method for streamflow gauging, demonstrating its wide applicability in diverse hydro-environmental scenarios.

Satellite Video Remote Sensing for Flood Model Validation

AbstractSatellite‐based optical video sensors are poised as the next frontier in remote sensing. Satellite video offers the unique advantage of capturing the transient dynamics of floods with the potential to supply hitherto unavailable data for the assessment of hydraulic models. A prerequisite for the successful application of hydraulic models is their proper calibration and validation. In this investigation, we validate 2D flood model predictions using satellite video‐derived flood extents and velocities. Hydraulic simulations of a flood event with a 5‐year return period (discharge of 722 m3 s−1) were conducted using Hydrologic Engineering Center—River Analysis System 2D in the Darling River at Tilpa, Australia. To extract flood extents from satellite video of the studied flood event, we use a hybrid transformer‐encoder, convolutional neural network (CNN)‐decoder deep neural network. We evaluate the influence of test‐time augmentation (TTA)—the application of transformations on test satellite video image ensembles, during deep neural network inference. We employ Large Scale Particle Image Velocimetry (LSPIV) for non‐contact‐based river surface velocity estimation from sequential satellite video frames. When validating hydraulic model simulations using deep neural network segmented flood extents, critical success index peaked at 94% with an average relative improvement of 9.5% when TTA was implemented. We show that TTA offers significant value in deep neural network‐based image segmentation, compensating for aleatoric uncertainties. The correlations between model predictions and LSPIV velocities were reasonable and averaged 0.78. Overall, our investigation demonstrates the potential of optical space‐based video sensors for validating flood models and studying flood dynamics.

Measuring Velocity and Discharge of High Turbidity Rivers Using an Improved Near-Field Remote-Sensing Measurement System

Large-scale particle image velocimetry (LSPIV) is a computer vision-based technique renowned for its precise and efficient measurement of river surface velocity. However, a crucial prerequisite for utilizing LSPIV involves camera calibration. Conventional techniques rely on ground control points, thus restricting their scope of application. This study introduced a near-field remote-sensing measurement system based on LSPIV, capable of accurately measuring river surface velocity sans reliance on ground control points. The system acquires gravity-acceleration data using a triaxial accelerometer and converts this data into a camera pose, thereby facilitating swift camera calibration. This study validates the system through method verification and field measurements. The method verification results indicate that the system’s method for retroactively deriving ground control-point coordinates achieves an accuracy exceeding 90%. Then, field measurements were performed five times to assess the surface velocity of the Datong River. These measured results were analyzed and compared with data collected from the radar wave velocity meter (RWCM) and the LS1206B velocity meter. Finally, a comprehensive sensitivity analysis of each parameter was conducted to identify those significantly impacting the river’s surface velocity. The findings revealed that this system achieved an accuracy exceeding 92% for all river surface velocities measured.

Satellite Video Remote Sensing for Estimation of River Discharge

AbstractWe demonstrate that river discharge can be estimated by deriving water surface velocity estimates from satellite‐derived video imagery when combined with high‐resolution topography of channel geometry. Large Scale Particle Image Velocimetry (LSPIV) was used to map surface velocity from 28 s of 5 Hz satellite video acquired at a 1.2 m nominal ground spacing over the Darling River, Tilpa, Australia, during a 1‐in‐5‐year flood. We stabilized and assessed the uncertainty of the residual motion induced by the satellite platform, enhancing our sub‐pixel motion analysis, and quantified the sensitivity of image extraction rates on computed velocities. In the absence of in situ observations, LSPIV velocity estimates were validated against predictions from a calibrated 2D hydrodynamic model. Despite the confounding influence of selecting a surface velocity depth‐averaging coefficient, inference of discharge was within 0.3%–15% compared with gauging station measurements. These results provide a valuable foundation for refining satellite video LSPIV techniques.

Comparative Assessment of Different Image Velocimetry Techniques for Measuring River Velocities Using Unmanned Aerial Vehicle Imagery

The accurate measurement of river velocity is essential due to its multifaceted significance. In response to this demand, remote measurement techniques have emerged, including large-scale particle image velocimetry (LSPIV), which can be implemented through cameras or unmanned aerial vehicles (UAVs). This study conducted water surface velocity measurements in the Xihu River, situated in Miaoli County, Taiwan. These measurements were subjected to analysis using five distinct algorithms (PIVlab, Fudaa-LSPIV, OpenPIV, KLT-IV, and STIV) and were compared with surface velocity radar (SVR) results. In the quest for identifying the optimal parameter configuration, it was found that an IA size of 32 pixels × 32 pixels, an image acquisition frequency of 12 frames per second (fps), and a pixel size of 20.5 mm/pixel consistently yielded the lowest values for mean error (ME) and root mean squared error (RMSE) in the performance of Fudaa-LSPIV. Among these algorithms, Fudaa-LSPIV consistently demonstrated the lowest mean error (ME) and root mean squared error (RMSE) values. Additionally, it exhibited the highest coefficient of determination (R2 = 0.8053). Subsequent investigations employing Fudaa-LSPIV delved into the impact of various water surface velocity calculation parameters. These experiments revealed that alterations in the size of the interrogation area (IA), image acquisition frequency, and pixel size significantly influenced water surface velocity. This parameter set was subsequently employed in an experiment exploring the incorporation of artificial particles in image velocimetry analysis. The results indicated that the introduction of artificial particles had a discernible impact on the calculation of surface water velocity. Inclusion of these artificial particles enhanced the capability of Fudaa-LSPIV to detect patterns on the water surface.

Evidence that superstructures comprise self-similar coherent motions in high Reynolds number boundary layers

We present experimental evidence that the superstructures in turbulent boundary layers comprise smaller, geometrically self-similar coherent motions. The evidence comes from identifying and analysing instantaneous superstructures from large-scale particle image velocimetry datasets acquired at high Reynolds numbers, capable of capturing streamwise elongated motions extending up to 12 times the boundary layer thickness. Given the challenge in identifying the constituent motions of the superstructures based on streamwise velocity signatures, a new approach is adopted that analyses the wall-normal velocity fluctuations within these very long motions, which reveals the constituent motions unambiguously. The conditional streamwise energy spectra of the Reynolds shear stress and the wall-normal fluctuations, corresponding exclusively to the superstructure region, are found to exhibit the well-known distance-from-the-wall scaling in the intermediate-scale range. It suggests that geometrically self-similar motions are the constituent motions of these very-large-scale structures. Investigation of the spatial organization of the wall-normal momentum-carrying eddies, within the superstructures, also lends empirical support to the concatenation hypothesis for the formation of these structures. The association between the superstructures and self-similar motions is reaffirmed on comparing the vertical coherence of the Reynolds-shear-stress-carrying motions, by computing conditionally averaged two-point correlations, which are found to match with the mean correlations. The mean vertical coherence of these motions, investigated for the log region across three decades of Reynolds numbers, exhibits a unique distance-from-the-wall scaling invariant with Reynolds number. The findings support modelling of these dynamically significant motions via data-driven coherent structure-based models.

Non-spectral linear depth inversion using drone-acquired wave field imagery

We present an efficient approach for video-based bathymetry survey. The proposed algorithm consists of preprocessing for drone-acquired video, extracting incident gravity wave, estimating celerity vector field, operating linear water depth inversion, and mapping bathymetry. The image preprocessing includes resizing, masking, barrel distortion correction, ortho-rectification, and georeferencing. The challenges posed by noisy, non-stationary, and irregular wave field data was resolved using ensemble empirical mode decomposition (EEMD), which is capable of extracting incident gravity wave structure without arbitrary initial guessing of frequencies. The extracted wave structures were used to estimate their celerity vector field using large-scale particle image velocimetry (LSPIV) technique. Given that wave celerity depends on water depth in nearshore based on linear dispersion relation, spatial distribution of water depths was determined from wave celerity estimates. The proposed algorithm was tested using the wave field videos and ground truth data, collected during the field experiment at Jangsa beach, located in the east coast of South Korea, and good agreement was achieved.

Multiple Camera Large-Scale Particle Image Velocimetry Feasibility for Rivers in Alaska

Multiple Camera Large-Scale Particle Image Velocimetry Feasibility for Rivers in Alaska

LSPIV measurements to assess the impact of a bridge on a weakly undulating flow

The modification of the surface velocity field of the weakly undulating flow in a reach of the Suquía River (Córdoba city, Argentina) due to the placement of a bridge was investigated in a physical model. Measurements using Large-Scale Particle Image Velocimetry (LSPIV) allowed evaluating how the hydraulic structure changes the size, shape, wavelength, and relative position to river centerline of the wave train downstream of the first ridge of the weak undular jump. Analysis and discussion of results in terms of the dimensionless parameters reported in the literature (obtained from laboratory studies) are offered. Results show that LSPIV is a reliable, fast low-cost tool to gather valuable information regarding the surficial flow field at the physical model scale in the presence of a weakly undular flow.

Aerodynamic benefits of drafting in speed skating: Estimates from in-field skater's wakes and wind tunnel measurements

The effect of drafting on the aerodynamic drag of a long-track speed skater is investigated in-field, at the 400m ice-rink of Thialf, Heerenveen. The Ring of Fire system is used to measure the flow downstream of an elite, isolated skater at approximately 11 m/s, transiting repeatedly through a tunnel filled with Helium-filled soap bubble flow tracers. Large-scale stereoscopic particle image velocimetry is used at an acquisition frequency of 500 Hz to obtain the near to far wake up to 11 m distance behind the skater. Over these 11 m, the center of gravity of the wake can shift up to 10 cm laterally, depending on the phase of the skating motion, and it moves about 15 cm to the floor. The former suggests that a trailing skater should slightly adapt its trajectory to achieve the lowest aerodynamic drag by drafting. The drag reduction of a trailing skater is estimated from the measurements on the isolated rider, assuming that the trailing rider's drag reduction only stems from the loss in total pressure in the wake of the first rider. The drag reduction is obtained with varying lateral and longitudinal distance between the leading and hypothetical trailing rider. It is observed that the peak reduction (∼40%) steeply decays with increasing lateral offset: at an offset of 50 cm the reduction is negligible. Instead, with increasing longitudinal offset, the decay is more gradual: at a distance of 11 m the reduction of 17% remains significant. The in-field estimations of the drag reduction are supported by wind tunnel measurements conducted on scaled skater models. Finally, the results obtained on the ice-rink indicate that a trailing skater should follow a slightly wider trajectory of about 20 cm, in comparison to the leading skater, to achieve the peak drag reduction during the entire skating stroke.

Space-Time Image Velocimetry Based on Improved MobileNetV2

Space-time image velocimetry (STIV) technology has achieved good performance in river surface-flow velocity measurement, but the application in a field environment is affected by bad weather or lighting conditions, which causes large measurement errors. To improve the measurement accuracy and robustness of STIV, we combined STIV with deep learning. Additionally, considering the light weight of the neural network model, we adopted MobileNetV2 and improved its classification accuracy. We name this method MobileNet-STIV. We also constructed a sample-enhanced mixed dataset for the first time, with 180 classes of images and 100 images per class to train our model, which resulted in a good performance. Compared to the current meter measurement results, the absolute error of the mean velocity was 0.02, the absolute error of the flow discharge was 1.71, the relative error of the mean velocity was 1.27%, and the relative error of the flow discharge was 1.15% in the comparative experiment. In the generalization performance experiment, the absolute error of the mean velocity was 0.03, the absolute error of the flow discharge was 0.27, the relative error of the mean velocity was 6.38%, and the relative error of the flow discharge was 5.92%. The results of both experiments demonstrate that our method is more accurate than the conventional STIV and large-scale particle image velocimetry (LSPIV).

River Surface Velocimetry Based on Virtual River Dataset and Modulated-RAFT

Many image-based velocimetry methods, such as large-scale particle image velocimetry (LSPIV), space-time image velocimetry (STIV) and optical flow velocimetry (OFV), have been proposed to estimate river surface velocity (RSV) efficiently and precisely. Among these methods, OFV-based methods have drawn great attention due to its high field resolution and low requirement of tracers. As a powerful approach for estimate optical flow accurately and efficiently, deep optical flow estimation is utilized in OFV-based methods as well. However, these methods often use irrelevant datasets for training due to the immeasurability of optical flow. Models obtained by such approaches suffer from limited generalization because of domain drift. Besides, high similarity of river surfaces often leads to ambiguous correlation volume extracted by deep optical flow estimation model, which can cause mismatch. This work proposed a method for accurately and robustly RSV estimation under velocity range of 0-6.0 m/s. Specifically, we introduced a method for generating relative optical flow datasets and proposed MRAFT, a deep optical flow estimation model combining with correlation volume modulation. Experiment results demonstrate that the generated virtual river datasets effectively improve the generalization of the model and MRAFT remarkably alleviates the mismatch. Our work facilitates the application of deep optical flow on RSV estimation and provides other related researches with optical flow datasets for fine-tuning.

Uncertainty Analysis for Image-Based Streamflow Measurement: The Influence of Ground Control Points

Large-scale particle image velocimetry (LSPIV) provides a cost-effective, rapid, and secure monitoring tool for streamflow measurements. However, surveys of ground control points (GCPs) might affect the camera parameters through the solution of collinearity equations and then impose uncertainty on the measurement results. In this paper, we explore and present an uncertainty analysis for image-based streamflow measurements with the main focus on the ground control points. The study area was Yufeng Creek, which is upstream of the Shimen Reservoir in Northern Taiwan. A monitoring system with dual cameras was set up on the platform of a gauge station to measure the surface velocity. To evaluate the feasibility and accuracy of image-based LSPIV, a comparison with the conventional measurement using a flow meter was conducted. Furthermore, the degree of uncertainty in LSPIV streamflow measurements influenced by the ground control points was quantified using Monte Carlo simulation (MCS). Different operations (with survey times from one to nine) and standard errors (30 mm, 10 mm, and 3 mm) during GCP measurements were considered. Overall, the impacts in the case of single GCP measurement are apparent, i.e., a shifted and wider confidence interval. This uncertainty can be alleviated if the coordinates of the control points are measured and averaged with three repetitions. In terms of the standard errors, the degrees of uncertainty (i.e., normalized confidence intervals) in the streamflow measurement were 20.7%, 12.8%, and 10.7%. Given a smaller SE in GCPs, less uncertain estimations of the river surface velocity and streamflow from LSPIV could be obtained.

Hydroinformatic tools and their potential in the search for missing persons in rivers

This research proposes a methodological approach to search floating bodies or floating body parts in rivers using hydroinformatic tools. These tools may allow possible search locations and potential release sites to be determined. The approach draws from hydrometric field information, two-dimensional (2D) hydrodynamic modeling, particle tracking (PT) models, and large-scale particle image velocimetry (LSPIV). This methodology was applied to a case study of the La Miel river in Colombia, where existing reports of deceased persons are available. A series of hydraulic accidents have been defined to represent the hydrodynamic and transport processes that occur in such situations. The results indicate that potential search locations, namely places where human floating bodies or floating body parts may be found, are principally on the left and right shores of recirculation systems.

LSPIV analysis of ship-induced wave wash

Ship-induced wave wash affects the hydromorphological and ecological state of rivers through various mechanisms. The direct proximity of the riverbank is usually the most exposed, as the hydrodynamic stresses are the highest in these shallow water areas. Contrary to the steady and almost still, natural flow conditions (i.e., no waves of anthropogenic source), shoaling and breaking of ship waves increase the hydrodynamic stresses by orders of magnitudes, having notable ecological consequences, and resulting in bank erosion as well. Due to the shallow water depths and temporary drying, conventional measurement techniques are no longer applicable in these areas. In this study, large-scale particle image velocimetry (LSPIV) is used to quantify the prevailing flow conditions. In the absence of ground truth data in the wave breaking region, a high-resolution computational fluid dynamics model—verified with field pressure and acoustic Doppler velocimetry data—is used for the cross-validation of the LSPIV results. The results underline the applicability of LSPIV for the hydrodynamic analysis of wave velocities in this special riverine swash zone, which is of key importance from the aspect of ecology and bank erosion as well.Graphical abstract

Comparison of dense optical flow and PIV techniques for mapping surface current flow in tidal stream energy sites

Marine renewable energy site and resource characterisation, in particular tidal stream energy, require detailed flow measurements which often rely on high-cost in situ instrumentation which is limited in spatial extent. We hypothesise uncrewed aerial vehicles (UAV) offer a low-cost and low-risk data collection method for tidal stream environments, as recently techniques have been developed to derive flow from optical videography. This may benefit tidal and floating renewable energy developments, providing additional insight into flow conditions and complement traditional instrumentation. Benefits to existing data collection methods include capturing flow over a large spatial extent synchronously, which could be used to analyse flow around structures or for site characterisation; however, uncertainty and method application to tidal energy sites is unclear. Here, two algorithms are tested: large-scale particle image velocimetry using PIVlab and dense optical flow. The methods are applied on video data collected at two tidal stream energy sites (Pentland Firth, Scotland, and Ramsey Sound, Wales) for a range of flow and environmental conditions. Although average validation measures were similar (~ 20–30% error), we recommend PIVlab processed velocity data at tidal energy sites because we find bias (underprediction) in optical flow for higher velocities (> 1 m/s).