Droplet Tracking Research Articles

AI-enabled droplet detection and tracking for agricultural spraying systems

This work leverages recent advancements in computer vision and deep learning to detect and track the motion of droplets captured by a camera. While classical computer vision techniques have been employed for detection and tracking, those approaches have limitations and are not trivially extended to droplets. We approach the problems of droplet detection and tracking through a data-driven framework, in which an annotated database of droplet images is built and object detection and tracking models are trained on this database. The accuracy of the model is evaluated and the whole process is discussed. At this point, droplet geometric properties can be extracted. This information is critical in understanding the effectiveness of a system that is spraying the droplets.

Quasi-3D visualization and local parameter measurement of horizontal and nearly horizontal liquid–liquid two-phase flows by using PLIF method

Horizontal and nearly horizontal liquid–liquid two-phase flows are widely encountered in petroleum, chemical, and other important industrial production fields. The three-dimensional (3D) flow visualization and the local parameter measurement are of great significance for understanding the flow dynamics and revealing the flow-pattern transition mechanism. In this paper, experiments of horizontal and nearly horizontal (+5°) liquid–liquid flow are carried out in an acrylic pipe with an inner diameter of 20 mm. Silicone oil (organic phase) and water/glycerol mixture (aqueous phase) are used as the test fluid. A novel planar laser-induced fluorescence (PLIF) system is designed to detect the flow structures. The incident direction of the laser sheet is along the radial section of the pipe, and the PLIF camera is positioned at 60° to the plane of the laser sheet, referred to as PLIF-60. In the experiment, the two-dimensional (2D) flow images at the pipe radial section are captured by the PLIF-60 system. Considering the effect of the angle between the laser sheet and the PLIF camera, the flow images are corrected to remove the horizontal and vertical distortions. The liquid–liquid stratified interface and the entrained droplets are identified in the corrected flow images. A droplet tracking method is proposed to derive the droplet diameter and its quantity distribution. Quasi-3D flow structures are reconstructed upon the accurate detection of the stratified interface and the entrained droplets. The evolution characteristics of the interface structures and the spatial distributions of droplets are studied. In addition, aqueous phase holdup and droplet entrainment ratios at pipe axial and radial sections are extracted based on the quasi-3D flow visualizations. The effects of the flow conditions and the pipe inclinations on the profiles of the aqueous phase holdup and the droplet entrainment ratio are uncovered.

Probing interdendritic flow and hot tearing during solidification using real time X-ray imaging and droplet tracking

Despite the well-known importance of controlling shrinkage-induced liquid flow in alloy castings to avoid the formation of catastrophic hot tears during the final stages of solidification, there has been little direct experimental measurement of liquid metal flow and hot tear formation under practical conditions. We use synchrotron X-rays to obtain radiographic video sequences of the solidification of monotectic Al-Pb alloys in which Pb droplets form as a fine-scale emulsion. We track and measure the velocity of thousands of Pb droplets as they move through interdendritic regions due to the effect of liquid to solid shrinkage during the final stages of solidification, up to the point of hot tear formation. Based on the droplet velocities, we present an analysis to estimate the interdendritic liquid velocity as solid fraction increases, and thus the shrinkage pressure drop driving the flow. The analysis is applied for video sequences obtained for both equiaxed and columnar microstructures, each under a range of cooling rates. Our measurements of the critical shrinkage-induced pressure for hot tear formation agree well with prior model-based and theoretical suggestions. The limitations and prospects for droplet tracking measurements of liquid metal flows are discussed.

DropTrack—Automatic droplet tracking with YOLOv5 and DeepSORT for microfluidic applications

Deep neural networks are rapidly emerging as data analysis tools, often outperforming the conventional techniques used in complex microfluidic systems. One fundamental analysis frequently desired in microfluidic experiments is counting and tracking the droplets. Specifically, droplet tracking in dense emulsions is challenging due to inherently small droplets moving in tightly packed configurations. Sometimes, the individual droplets in these dense clusters are hard to resolve, even for a human observer. Here, two deep learning-based cutting-edge algorithms for object detection [you only look once (YOLO)] and object tracking (DeepSORT) are combined into a single image analysis tool, DropTrack, to track droplets in the microfluidic experiments. DropTrack analyzes input microfluidic experimental videos, extracts droplets' trajectories, and infers other observables of interest, such as droplet numbers. Training an object detector network for droplet recognition with manually annotated images is a labor-intensive task and a persistent bottleneck. In this work, this problem is partly resolved by training many object detector networks (YOLOv5) with several hybrid datasets containing real and synthetic images. We present an analysis of a double emulsion experiment as a case study to measure DropTrack's performance. For our test case, the YOLO network trained by combining 40% real images and 60% synthetic images yields the best accuracy in droplet detection and droplet counting in real experimental videos. Also, this strategy reduces labor-intensive image annotation work by 60%. DropTrack's performance is measured in terms of mean average precision of droplet detection, mean squared error in counting the droplets, and image analysis speed for inferring droplets' trajectories. The fastest configuration of DropTrack can detect and track the droplets at approximately 30 frames per second, well within the standards for a real-time image analysis.

One-Step Digital Droplet Auto-Catalytic Nucleic Acid Amplification with High-Throughput Fluorescence Imaging and Droplet Tracking Computation.

Digital droplet technology has emerged as a powerful new tool for biomarker analysis. Temperature cycling, enzymes, and off-chip processes are, nevertheless, always required. Herein, we constructed a digital droplet auto-catalytic hairpin assembly (ddaCHA) microfluidic system to achieve digital quantification of single-molecule microRNA (miRNA). The designed continuous chip integrates droplet generation, incubation, and fluorescence imaging on the chip, avoiding the requirement for extra droplet re-collection and heating operations. Clearly, the digital readout was obtained by partitioning miRNA into many individual pL-sized small droplets in which the target molecule is either present ("positive") or absent ("negative"). Importantly, the suggested enzyme-free auto-catalytic hairpin assembly (aCHA) in droplets successfully mitigated the effects of the external environment and thermal cycling on droplets, and its reaction rate is significantly superior to that of traditional CHA. We got excellent sensitivity with a linear correlation from 1 pM to 10 nM and a detection limit of 0.34 pM in the fluorescence spectrum section, as well as high selectivity to other miRNAs. Furthermore, the minimum target concentration could be reduced to 10 fM based on the high-throughput tracking computation of fluorescent droplets with a self-developed Python script, and the fluorescence intensity distribution agreed well with the theoretical value, demonstrating that it is feasible to detect miRNA efficiently and accurately, which has great potential applications in clinical diagnostics and biochemical research.

Specific lipid droplet imaging of atherosclerotic plaques and fatty liver using an imidazole-based fluorescence probe

Selective labelling and high-fidelity tracking of intracellular lipid droplets (LDs) can provide direct insight into pathological processes associated with abnormal lipid metabolism and therefore are of great value. We herein present a novel imidazole-based probe, TITM, which shows the aggregation-induced emission property and a remarkable fluorescence increase in sunflower oil than in water. With the advantages of good lipophilicity, biocompatibility, and LDs-targeted specificity, TITM allows the real-time and long-term tracking of dynamic LDs in living cells with high-resolution imaging performance. By establishment of LDs-related atherosclerosis (AS) and fatty liver mouse model, the imaging performance of TITM at pathological tissue level is well studied to show impressive selectivity and resolution.

Structure of the ethanol spray flame under conventional and MILD conditions

A numerical examination is carried out on the experimental MILD turbulent spray burner of Delft Spray in Hot Co-flow, DSHC. The RANS approach is combined with the Lagrangian tracking of droplets for modeling the pressure-swirl atomizer. The EDC combustion model is employed, to take into account detailed chemical mechanisms. Predictions show that the structure of the triple-flame of liquid fuel in the atmospheric condition is different from the conventional gaseous triple-flames. Furthermore, the structure of the MILD spray flame is compared to the non-MILD spray flame. The flame index shows that the triple-flame structure in the conventional spray flame that consists of two outer and middle diffusion branches, and a premixed inner branch converts to a double-flame structure under MILD condition, including an outer diffusion and a premixed inner branch. In another step, sensitivity analysis relative to the temperature, and oxygen concentration of the co-flow stream is performed independently, which was not possible in the experiment. The parameter of ignition delay time is used for the selection of parametric study cases. Moderate preheating could lead to deviation from the MILD combustion regime, while high preheating is beneficial. NO pollutant generation shows more sensitivity to dilution rather than preheating in MILD spray flame. This study showed that increasing or decreasing each of the temperature and oxygen concentration doesn’t improve all the effective aspects toward achieving the MILD condition. So, the effective parameters should be changed according to a specific and more precise goal.

Ultrabright and Highly Polarity‐Sensitive NIR‐I/NIR‐II Fluorophores for the Tracking of Lipid Droplets and Staging of Fatty Liver Disease (Adv. Funct. Mater. 12/2022)

Fatty Liver Detection In article number 2109929, Fuyou Li, Yuling Qin, Li Wu, and co-workers demonstrate an effective strategy to construct a NIR-I/NIR-II fluorescent probe with a highly sensitive response to environment polarity, ultrabright emission, and superior photostability for specific intracellular lipid droplet imaging in living cells and in vivo. The findings promote the potential of fluorescence imaging for realtime fatty liver disease analysis in vivo.

Engineering BODIPY-based near-infrared nanoparticles with large Stokes shifts and aggregation-induced emission characteristics for organelle specific bioimaging.

Lipid droplets (LDs) and lysosomes, as two important subcellular organelles, play specific and indispensable roles in various cellular processes. The development of efficient LD- and lysosome-specific fluorescent bio-probes is of great importance. However, current commercial lipid droplet- (LD) and lysosome-specific fluorescent specific bio-probes often suffer from the aggregation-caused quenching (ACQ) effect, short absorption and emission wavelengths, poor photostability and low specificity. Herein, a typical ACQ luminogen BODIPY was directly conjugated to strong electron donating triarylamine units at its α-positions, giving near-infrared (NIR) fluorescent materials TPAB and 2TPAB with aggregation-induced emission (AIE). Both TPAB and 2TPAB nanoparticles were obtained by self-assembly, and showed NIR emissions, large Stokes shifts, good photostability and two-photon absorption. These nanoparticles presented remarkable bioimaging performances and were shown to specifically localize in LDs or lysosomes, respectively, depending on the number of triarylamine units attached. They have been successfully used to detect endogenous LD overproduction, and monitor abnormal activities of LDs/lysosomes, as well as real-time track the lipophagy process in cells. Their far NIR emission and two-photon excitation further supported their promising bioimaging application for lipid droplet tracking in liver tissue and live zebrafish larvae. Our work here enriches BODIPY based NIR AIE dyes and provides organelle specific bio-probes which are superior to currently used commercial ones.

Ultrabright and Highly Polarity‐Sensitive NIR‐I/NIR‐II Fluorophores for the Tracking of Lipid Droplets and Staging of Fatty Liver Disease

AbstractVisualizing the progression of fatty liver disease in vivo is of great significance in biomedical research of fatty liver, but hindered by the lack of suitable approaches with deep penetration, high specificity, and real‐time observability for fatty liver imaging. Herein, two bright NIR‐I/NIR‐II fluorophores with D–π–A–π–D structures are designed and synthesized for fatty liver tracking in vivo. Such fluorophores show an ultrahigh quantum yield of 49.2% and 17.3% at the emission maximum of 930 and 975 nm, respectively. Most interestingly, these two fluorophores display ultrasensitive fluorescence response (over 100‐fold enhancement) toward the variations of environment polarity. Accordingly, further doping of these fluorophores into amphiphilic organic matrix allows the formation of ultrabright NIR‐I/NIR‐II fluorescent nanoprobes for high‐resolution fluorescence imaging of the vessels. Additionally, with the assistance of such nanoprobes and NIR‐II fluorescence imaging, real‐time monitoring of heartbeat rate and the respiratory rate, accurate identification, and imaging‐guided surgery of the submillimeter scale lymph nodes are successfully realized. Moreover, the as‐designed fluorophores show superior performance for indicating lipid droplets (LDs) in living cells. Ultimately, precise monitoring of fatty liver with NIR‐I/NIR‐II fluorescent LDs probe is first confirmed, demonstrating the possibility for staging fatty liver disease with a contact‐free method.

High-performance CFD for Respiratory Droplet Turbulent Dispersion in a Ventilated City Bus

This work focuses on the development of a simulation strategy able to quantify risks of airborne virus contagion in many scenarios found in enclosed domains by using high-fidelity fluid dynamics simulation to predict the trajectories and distribution of virus-loaded respiratory droplets over long times. Large-Eddy simulation is used to predict the turbulent flow fields in a city bus for different operating conditions of the Air Conditioning system. The time-averaged velocity distributions and associated turbulent kinetic energy are shown to be drastically dependent on the studied operating conditions. Lagrangian tracking of respiratory droplets is then used over long times on statically converged Eulerian flow fields to investigate their turbulent dispersion depending on the emitter position in the bus. Importance of air conditioning conditions on respiratory droplet trajectories and concentration in the configuration is illustrated indicating that air treatment devices play a crucial role in the mitigation solution of airborne virus propagation.

Construction of a robust polarity sensitive platform and its application for tracking of lipid droplets decrease under oxidative stress in live cells

There are many hydrophilic and hydrophobic organelles and regions in cells, which play important roles in the physiological activities of cells. In this work, we have developed a new polarity sensitive fluorescent platform (DN-NR) to detect cell polarity and lipid droplets (LDs) evolution. DN-NR is composed of the fluorophores dansyl chloride (DN) and a nile red (NR) derivative. DN-NR can detect polarity changes in both green and red detection channels. Moreover, by taking advantage of solvatochromism and the aggregation-caused fluorescence quenching (ACQ) effect of chromophores, DN-NR responds to polarity changes via emission changes in both hydrophobic and hydrophilic environments. DN-NR possesses a strong positive solvatochromism with solvent polarity change, and the response mechanism was further investigated by lifetime measurement and DFT calculation. As proved by co-staining experiments, DN-NR can effectively detect LDs in living HeLa cells, and the LDs evolution was dynamically visualized. Experimental results indicated that oxidative stress induced by H2O2 treatment leads to a sharp decrease of LDs. Therefore, DN-NR has broad prospective applications for biomedical research and diagnosis in LDs tracking and analysis.

Intelligent droplet tracking with correlation filters for digital microfluidics

Tracking the movement of droplets in digital microfluidics is essential to improve its control stability and obtain dynamic information for its applications such as point-of-care testing, environment monitoring and chemical synthesis. Herein, an intelligent, accurate and fast droplet tracking method based on machine vision is developed for applications of digital microfluidics. To continuously recognize the transparent droplets in real-time and avoid the interferes from background patterns or inhomogeneous illumination, we introduced the correlation filter tracker, enabling online learning of the multi-features of the droplets in Fourier domain. Results show the proposed droplet tracking method could accurately locate the droplets. We also demonstrated the capacity of the proposed method for estimation of the droplet velocity as faster as 20 mm/s, and its application in online monitoring the Griess reaction for both colorimetric assay of nitrite and study of reaction kinetics.

High-speed digital off-axis holography to study atomization and evaporation of burning droplets

Droplet formation and evaporation are key processes of spray combustion, the mechanism of which is yet not fully understood. In this study, we develop a method based on high-speed (25 kHz) digital off-axis holography (DOH) to visualize and measure burning ethanol droplets in the atomization. Ligaments and fragments from a burning drop internally blown to disintegrate are studied after hologram processing. Reconstructed images are nearly free from influences of the flame with the measured droplets showing clear borders. Size and depth location uncertainties reduce to nearly 1/10 of those of commonly used with digital in-line holography. A time-resolved observation shows the dynamic formation of secondary droplets with stages classified as vapor penetration, sheet breakup, and ligament disintegration. Sizes and velocities of fragments are extracted and their distributions are analyzed at 1.12 ms, 3.12 ms, and 13 ms after the initiation of atomization. Results show that new secondary droplets have larger sizes and smaller velocities with time. Lagrangian measurements are performed with 3D droplet tracking algorithm. Evaporation rates of 31 droplets with different positions, sizes, and velocities are obtained and the results are in good accordance with previous studies. The DOH technique is promising to investigate the combustion of individual fuel droplets in a polydisperse droplet cloud as the real spray combustion.

Role of pulsatility on particle dispersion in expiratory flows.

Expiratory events, such as coughs, are often pulsatile in nature and result in vortical flow structures that transport respiratory particles. In this work, direct numerical simulation (DNS) of turbulent pulsatile jets, coupled with Lagrangian particle tracking of micron-sized droplets, is performed to investigate the role of secondary and tertiary expulsions on particle dispersion and penetration. Fully developed turbulence obtained from DNS of a turbulent pipe flow is provided at the jet orifice. The volumetric flow rate at the orifice is modulated in time according to a damped sine wave, thereby allowing for control of the number of pulses, duration, and peak amplitude. Thermodynamic effects, such as evaporation and buoyancy, are neglected in order to isolate the role of pulsatility on particle dispersion. The resulting vortex structures are analyzed for single-, two-, and three-pulse jets. The evolution of the particle cloud is then compared to existing single-pulse models. Particle dispersion and penetration of the entire cloud are found to be hindered by increased pulsatility. However, the penetration of particles emanating from a secondary or tertiary expulsion is enhanced due to acceleration downstream by vortex structures.

High frame rate emission spectroscopy for ablation tests in plasma wind tunnel.

This article describes a novel high frame rate emission spectroscopy setup developed for measurements in high enthalpy flow fields. The optical setup and the associated hardware arrangements are described in detail followed by test case data to demonstrate the capability of recording spectral images at 1 kHz frame rate. The new system is based on a classical Czerny-Turner spectrograph but with a particular setup for high frame rate detection using a Generation II intensifier coupled with a high-speed camera. The high frame rate spectral images acquired enable, for the first time, investigation of the spatial distribution and temporal tracking and evolution of molten droplets of an ablating sample. In this paper, an example is shown from ablating meteorite samples tested in a high enthalpy plasma flow field corresponding to a flight scenario at an altitude of 80 km. This new instrumental configuration allows emission spectroscopic analysis of transient phenomena simulated in the high enthalpy ground test facilities with kHz resolution. The particular feature of this system is the ability to measure very faint spectral lines at high temporal and spatial resolution.

A family of lipotropic AIEgens for high-fidelity dynamic tracking of lipid droplets in living cells

As dynamic subcellular organelles, lipid droplets (LDs) are associated with energy homeostasis and handling of cellular lipids and proteins, and their abnormal activity and quantity are involved in many serious diseases, even cancer. Thus, the biological morphology of LDs monitoring is very important to disease screening and biomedical research. In this work, we developed a series of aggregation-induced emission luminogens (AIEgens, TFI, MTFI and MOTFI) via very ordinary synthetic technology and convenient purification. Owing to large Stokes shifts (greater than 140 nm), bright red emissions, excellent photostability and good biocompatibility, these lipotropic AIEgens can specially accumulate in LDs and visually ‘light up’ them. In addition, the MTFI was successfully engaged to high-fidelity monitoring and tracking the LDs' movements in living cells. Moreover, the LO2 and HepG2 cells could be clearly distinguished according to the significantly different fluorescence intensity of MTFI marked LDs. Thus these AIEgens may become promising imaging molecule tools to investigate the physiological and pathology function of intracellular LDs, and show great potential for diagnosis of LDs-related diseases.

Motion tracking and analysis of spray water droplets studied by high-speed photography using an iPhone X

Smartphones are not only becoming an inseparable part of our daily lives, but also a low-cost, powerful optical imaging tool for more and more scientific research applications. In this work, we hypothesized that some smartphones could be used as a low-cost, high-speed, photographic alternative to expensive equipment, such as those typically found in scientific research labs, to accurately perform motion tracking and analysis of fast-moving objects. This demonstrative experiment was done at home at very low cost by using a low-power green laser (<0.1 W) purchased from Amazon, an iPhone X, and a water spray bottle. The sprayed water droplets were illuminated by a divergent green laser beam and recorded with a smartphone using the Slow-Motion Video function (240 frames per second), which allows us to record high-speed events. The Slow-Motion Video was converted to a series of consecutive images taken every 1/240 second (or 4.16 ms), enabling tracking and motion analysis of the sprayed water droplets. By analyzing these consecutive images, the speed and flight trajectory of water droplets in the air were obtained, thereby enabling us to estimate the area of the water droplets landing on the ground.

A fluorogenic probe for dynamic tracking of lipid droplets’ polarity during the evolution of cancer

Exploring the changes in the polarity of intracellular lipid droplets (LDs) during the evolution of cancer is important for cancer detection and treatment.

Patterning of colloidal droplet deposits on soft materials

Abstract