Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Critical Step”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 235 records · Page 13
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

ORNL Infrastructure: Status Update on BEA Research Reactor Cask Planning Activities Completed During FY 2025

This report documents Oak Ridge National Laboratory’s (ORNL’s) FY 2025 progress toward establishing readiness to use the Battelle Energy Alliance (BEA) Research Reactor (BRR) shipping cask in support of the Nuclear Science User Facilities (NSUF). The BRR cask will provide a new shared infrastructure capability for transporting irradiated fuels and materials between ORNL and Idaho National Laboratory (INL), thereby supporting NSUF’s mission of enabling user access to advanced nuclear research facilities. In FY 2025, ORNL advanced the regulatory, contractual, safety review, and planning activities necessary to qualify its facilities and staff for BRR cask handling. Although loaded shipments were delayed due to fabrication lead times for Orano Federal Services LLC’s internal basket hardware, the program progressed to the point where an empty cask dry run is scheduled for October 2025. This dry run represents a critical step in demonstrating ORNL’s ability to receive, handle, and return the BRR cask.

99 GENERAL AND MISCELLANEOUS↗
DOE OSTI2024

Dataset for "Enrichment of phosphates, lead, and mixed soil-organic particles in INPs at the Southern Great Plains site"

Ice nucleating particles (INPs) are rare particles that initiate primary ice formation, a critical step required for subsequent important cloud microphysical processes that ultimately govern cloud phase and cloud radiative properties. Laboratory studies have found that organic-rich dusts, such as those found in soils, are more efficient INPs compared to mineral dust. However, the atmospheric relevance of these organic-rich dusts are not well understood, particularly in regions with significant agricultural activity. The Agricultural Ice nuclei at the Southern Great Plains field campaign (AGINSGP) was conducted in rural Oklahoma to investigate how soil dusts contribute to INP populations in the Great Plains. We present chemical characterization of ambient and ice crystal residual particles from a single day of sampling using single particle mass spectrometry (SPMS) and scanning microscopy. Ambient particle concentrations were primarily carbonaceous or secondary aerosol, while dust particles were enhanced in the residual particles. Dust particles measured during residual sampling contained elevated signals for phosphate ($^{63}$PO$_2^-$ and $^{79}$PO$_3^-$) and lead ($^{208}$Pb$_2^+$). Nitrate was slightly depleted relative to ambient dust, and strong sulfate signals were not seen in the residual particles measured by the miniSPLAT. The residual and ambient particles measured by the miniSPLAT were approximately the same size, though this may be a reflection of the SPMS size-dependent transmission efficiency and not that of physical truth. This study shows that organic-rich soils may be important contributors to the ambient INP population in agricultural regions.

Cornwell, Gavin C↗
DOE OSTI2026

Lignin Extraction and Condensation as a Function of Temperature, Residence Time, and Solvent System in Flow-through Reactors

Solvolytic extraction of lignin from biomass is a critical step in lignin-first biorefining, including the reductive catalytic fractionation (RCF) process. Key to optimal RCF processing is the ability to rapidly extract lignin from biomass at high delignification extents and transfer the lignin molecules to a catalyst surface in a time frame that minimizes lignin condensation reactions. Here, we use a flow-through reactor to study the effects of temperature (175–250 °C), residence time (9 to 36 min), and solvent composition (methanol and methanol–water) on lignin extraction and condensation. We evaluated three metrics at each condition: total delignification, delignification rate, and extent of condensation, the latter measured by a decrease in monomer yield for batch hydrogenolysis reactions of solvolysis liquor compared to batch RCF reactions. We observe that delignification is predominantly determined by temperature, while residence time dictates the lignin condensation extent. Moreover, the extent of both extraction and condensation increased in the methanol–water solvent system compared to that in the methanol system. Lignin extracted in methanol is stable up to 18-min residence times at or below 225 °C, while a majority of the lignin extracted in methanol–water is condensed with a 9-min residence time at 200 °C. These results can inform reactor designs and solvent selection for lignin-first biorefining processes that aim to physically separate the biomass and catalyst.

biorefining↗
DOE OSTI2023

Dynamic Network Analysis of Nuclear Science Literature for Research Influence Assessment

Analyzing nuclear science literature via data-driven methods is a critical step for assessing research influence and technology advancements. Indicators of scholarly activities may be buried in large volumes of nuclear research publications and collaboration networks over time. Mining for relevant scholarly influence trends in large volumes of text can be computationally challenging; however, open-source information on research collaborations over time can offer opportunities to extract meaningful insights. While network centrality analysis of scholarly research provides topology-based insights, additional emphasis on dynamics associated with the diffusion of information through these networks is important. Here this paper represents a step in that direction through the development of a novel dynamic network analysis framework and computational engine to identify key entities and capabilities over time within global scholarly nuclear science collaboration networks. Network theoretic, stochastic simulation, and optimization methods are leveraged to address variability in scholarly interactions, influence propagation, and collaboration patterns via network connections. A topic-aware influence maximization algorithm is developed to address the goal of identifying key influential authors in diverse research topics over time. Efficient parallelized implementation of the algorithm is applied to reduce computational costs. A proof-of-concept case study using open-source Scopus data with 33,517 published nuclear research papers from 2000-2019 is presented and representative analytic insights are generated. Broad implications of these insights are discussed and future research directions are also identified.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗
DOE OSTI2020

Stability of Floodplain Subsurface Microbial Communities Through Seasonal Hydrological and Geochemical Cycles

Riparian floodplains represent an interaction zone between the terrestrial subsurface and rivers, where regional groundwater flows, infiltration, and evapotranspiration drive mixing of water and import/export of nutrients and contaminants. These dynamics create seasonally transient redox conditions that drive biogeochemical transformations, which strongly modify groundwater quality. Microbial responses to changing hydrological conditions are perhaps the critical step connecting hydrology to geochemical transformations and groundwater quality, yet are not well understood. We aimed to address this knowledge gap by monitoring seasonal transitions at the U.S. Department of Energy legacy uranium ore processing site in Riverton, WY, through spring-summer-fall hydrological transitions. Our goal was to characterize the microbial community throughout the soil profile, down to the saturated aquifer, and observe its response to wet-dry transitions across a full season and compare to changes in geochemistry and hydrology. Next-generation sequencing was employed to identify biogeochemically-relevant microbial taxa based on the 16S rRNA gene; we found a broad diversity of microbial clades including taxa involved in sulfur and metal cycling, as well as nitrification. These data were paired with measurements of soil moisture, major nutrients and cations, and trace elements. Overall microbial community composition was dependent on soil depth or type, with seasonal effects only observed in the topsoil or subsurface aquifer. This finding indicates that microbial communities in the transiently-reduced center of the soil profile at the Riverton, WY site are remarkably stable, despite moisture and redox inversions. In addition, these communities likely impact the communities in surrounding soil horizons through export of metabolites and solutes as the water table rises and falls throughout the season.

16S rRNA↗
DOE OSTI2026

Beyond microbial abundance: metadata integration enhances disease prediction in human microbiome studies

Multiple studies have highlighted the interaction of the human microbiome with physiological systems such as the gut, immune, liver, and skin, via key axes. Advances in sequencing technologies and high-performance computing have enabled the analysis of large-scale metagenomic data, facilitating the use of machine learning to predict disease likelihood from microbiome profiles. However, challenges such as compositionality, high dimensionality, sparsity, and limited sample sizes have hindered the development of actionable models. One strategy to improve these models is by incorporating key metadata from both the human host and sample collection/processing protocols. This remains challenging due to sparsity and inconsistency in metadata annotation and availability. In this paper, we introduce a machine learning-based pipeline for predicting human disease states by integrating host and protocol metadata with microbiome abundance profiles from 68 different studies, processed through a consistent pipeline. Our findings indicate that metadata can enhance machine learning predictions, particularly at higher taxonomic ranks like Kingdom and Phylum, though this effect diminishes at lower ranks. Our study leverages a large collection of microbiome datasets comprising 11,208 samples, therefore enhancing the robustness and statistical confidence of our findings. This work is a critical step toward utilizing microbiome and metadata for predicting diseases such as gastrointestinal infections, diabetes, cancer, and neurological disorders.

Mathematics and Computing↗
DOE OSTI2022

Fast Characterization of Inducible Regions of Atrial Fibrillation Models With Multi-Fidelity Gaussian Process Classification

Computational models of atrial fibrillation have successfully been used to predict optimal ablation sites. A critical step to assess the effect of an ablation pattern is to pace the model from different, potentially random, locations to determine whether arrhythmias can be induced in the atria. In this work, we propose to use multi-fidelity Gaussian process classification on Riemannian manifolds to efficiently determine the regions in the atria where arrhythmias are inducible. We build a probabilistic classifier that operates directly on the atrial surface. We take advantage of lower resolution models to explore the atrial surface and combine seamlessly with high-resolution models to identify regions of inducibility. We test our methodology in 9 different cases, with different levels of fibrosis and ablation treatments, totalling 1,800 high resolution and 900 low resolution simulations of atrial fibrillation. When trained with 40 samples, our multi-fidelity classifier that combines low and high resolution models, shows a balanced accuracy that is, on average, 5.7% higher than a nearest neighbor classifier. We hope that this new technique will allow faster and more precise clinical applications of computational models for atrial fibrillation. All data and code accompanying this manuscript will be made publicly available at: https://github.com/fsahli/AtrialMFclass.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Enzyme Complexes of Ptr4CL and PtrHCT Modulate Co-enzyme A Ligation of Hydroxycinnamic Acids for Monolignol Biosynthesis in Populus trichocarpa

Co-enzyme A (CoA) ligation of hydroxycinnamic acids by 4-coumaric acid:CoA ligase (4CL) is a critical step in the biosynthesis of monolignols. Perturbation of 4CL activity significantly impacts the lignin content of diverse plant species. In Populus trichocarpa, two well-studied xylem-specific Ptr4CLs (Ptr4CL3 and Ptr4CL5) catalyze the CoA ligation of 4-coumaric acid to 4-coumaroyl-CoA and caffeic acid to caffeoyl-CoA. Subsequently, two 4-hydroxycinnamoyl-CoA:shikimic acid hydroxycinnamoyl transferases (PtrHCT1 and PtrHCT6) mediate the conversion of 4-coumaroyl-CoA to caffeoyl-CoA. Here, we show that the CoA ligation of 4-coumaric and caffeic acids is modulated by Ptr4CL/PtrHCT protein complexes. Downregulation of PtrHCTs reduced Ptr4CL activities in the stem-differentiating xylem (SDX) of transgenic P. trichocarpa. The Ptr4CL/PtrHCT interactions were then validated in vivo using biomolecular fluorescence complementation (BiFC) and protein pull-down assays in P. trichocarpa SDX extracts. Enzyme activity assays using recombinant proteins of Ptr4CL and PtrHCT showed elevated CoA ligation activity for Ptr4CL when supplemented with PtrHCT. Numerical analyses based on an evolutionary computation of the CoA ligation activity estimated the stoichiometry of the protein complex to consist of one Ptr4CL and two PtrHCTs, which was experimentally confirmed by chemical cross-linking using SDX plant protein extracts and recombinant proteins. Based on these results, we propose that Ptr4CL/PtrHCT complexes modulate the metabolic flux of CoA ligation for monolignol biosynthesis during wood formation in P. trichocarpa.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Classification of Intensity Distributions of Transmission Eigenchannels of Disordered Nanophotonic Structures Using Machine Learning

Light-matter interaction optimization in complex nanophotonic structures is a critical step towards the tailored performance of photonic devices. The increasing complexity of such systems requires new optimization strategies beyond intuitive methods. For example, in disordered photonic structures, the spatial distribution of energy densities has large random fluctuations due to the interference of multiply scattered electromagnetic waves, even though the statistically averaged spatial profiles of the transmission eigenchannels are universal. Classification of these eigenchannels for a single configuration based on visualization of intensity distributions is difficult. However, successful classification could provide vital information about disordered nanophotonic structures. Emerging methods in machine learning have enabled new investigations into optimized photonic structures. In this work, we combine intensity distributions of the transmission eigenchannels and the transmitted speckle-like intensity patterns to classify the eigenchannels of a single configuration of disordered photonic structures using machine learning techniques. Specifically, we leverage supervised learning methods, such as decision trees and fully connected neural networks, to achieve classification of these transmission eigenchannels based on their intensity distributions with an accuracy greater than 99%, even with a dataset including photonic devices of various disorder strengths. Simultaneous classification of the transmission eigenchannels and the relative disorder strength of the nanophotonic structure is also possible. Our results open new directions for machine learning assisted speckle-based metrology and demonstrate a novel approach to classifying nanophotonic structures based on their electromagnetic field distributions. These insights can be of paramount importance for optimizing light-matter interactions at the nanoscale.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Implementation of a Practical Teaching Course on Protein Engineering

Protein Engineering is a highly evolved field of engineering aimed at developing proteins for specific industrial, medical, and research applications. Here, we present a practical teaching course to demonstrate fundamental techniques used to express, purify and analyze a recombinant protein produced in Escherichia coli—the enhanced green fluorescent protein (eGFP). The methodologies used for eGFP production were introduced sequentially over six laboratory sessions and included (i) bacterial growth, (ii) sonication (for cell lysis), (iii) affinity chromatography and dialysis (for eGFP purification), (iv) bicinchoninic acid (BCA) and fluorometry assays for total protein and eGFP quantification, respectively, and (v) sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) for qualitative analysis. All groups were able to isolate the eGFP from the cell lysate with purity levels up to 72%. Additionally, a mass balance analysis performed by the students showed that eGFP yields up to 46% were achieved at the end of the purification process following the adopted procedures. A sensitivity analysis was performed to pinpoint the most critical steps of the downstream processing.

Gomes, Luciana (ORCID:0000000289921097)↗
DOE OSTI2023

Using Ultrasound Image Augmentation and Ensemble Predictions to Prevent Machine-Learning Model Overfitting

Deep learning predictive models have the potential to simplify and automate medical imaging diagnostics by lowering the skill threshold for image interpretation. However, this requires predictive models that are generalized to handle subject variability as seen clinically. Here, we highlight methods to improve test accuracy of an image classifier model for shrapnel identification using tissue phantom image sets. Using a previously developed image classifier neural network—termed ShrapML—blind test accuracy was less than 70% and was variable depending on the training/test data setup, as determined by a leave one subject out (LOSO) holdout methodology. Introduction of affine transformations for image augmentation or MixUp methodologies to generate additional training sets improved model performance and overall accuracy improved to 75%. Further improvements were made by aggregating predictions across five LOSO holdouts. This was done by bagging confidences or predictions from all LOSOs or the top-3 LOSO confidence models for each image prediction. Top-3 LOSO confidence bagging performed best, with test accuracy improved to greater than 85% accuracy for two different blind tissue phantoms. This was confirmed by gradient-weighted class activation mapping to highlight that the image classifier was tracking shrapnel in the image sets. Overall, data augmentation and ensemble prediction approaches were suitable for creating more generalized predictive models for ultrasound image analysis, a critical step for real-time diagnostic deployment.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

Optimize Electron Beam Energy toward In Situ Imaging of Thick Frozen Bio-Samples with Nanometer Resolution Using MeV-STEM

To optimize electron energy for in situ imaging of large biological samples up to 10 μm in thickness with nanoscale resolutions, we implemented an analytical model based on elastic and inelastic characteristic angles. This model has been benchmarked by Monte Carlo simulations and can be used to predict the transverse beam size broadening as a function of electron energy while the probe beam traverses through the sample. As a result, the optimal choice of the electron beam energy can be realized. In addition, the impact of the dose-limited resolution was analysed. While the sample thickness is less than 10 μm, there exists an optimal electron beam energy below 10 MeV regarding a specific sample thickness. However, for samples thicker than 10 μm, the optimal beam energy is 10 MeV or higher depending on the sample thickness, and the ultimate resolution could become worse with the increase in the sample thickness. Moreover, a MeV-STEM column based on a two-stage lens system can be applied to reduce the beam size from one micron at aperture to one nanometre at the sample with the energy tuning range from 3 to 10 MeV. In conjunction with the state-of-the-art ultralow emittance electron source that we recently implemented, the maximum size of an electron beam when it traverses through an up to 10 μm thick bio-sample can be kept less than 10 nm . This is a critical step toward the in situ imaging of large, thick biological samples with nanometer resolution.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Enabling Industrial Re-Use of Large-Format Additive Manufacturing Molding and Tooling

Large-format additive manufacturing (LFAM) is an enabling manufacturing technology capable of producing large parts with highly complex geometries for a wide variety of applications, including automotive, infrastructure/construction, and aerospace mold and tooling. In the past decade, the LFAM industry has seen widespread use of bio-based, glass, and/or carbon fiber reinforced thermoplastic composites which, when printed, serve as a lower-cost alternative to metallic parts. One of the highest-volume materials utilized by the industry is carbon fiber (CF)-filled polycarbonate (PC), which in out-of-autoclave applications can achieve comparable mechanical performance to metal at a significantly lower cost. Previous work has shown that if this material is recovered at various points throughout the manufacturing process for both the lab and pilot scale, it can be mechanically recycled with minimal impacts on the functional performance and printability of the material while significantly reducing the feedstock costs. End-of-life (EOL) CF-PC components were processed through industrial shredding, melt compounding, and LFAM equipment, followed by evaluation of the second-life material properties. Experimental assessments included quantitative analysis of fiber length attrition, polymer molecular weight degradation using gel permeation chromatography (GPC), density changes via pycnometry, thermal performance using dynamic mechanical analysis (DMA), and mechanical performance (tensile properties) in both the X- and Z-directions. Results demonstrated a 24.6% reduction in average fiber length compared to virgin prints, accompanied by a 21% decrease in X-direction tensile strength and a 39% reduction in tensile modulus. Despite these reductions, Z-direction tensile modulus improved by 4%, density increased by 6.8%, and heat deflection temperature (HDT) under high stress retained over 97% of its original value. These findings underscore the potential for integrating mechanically recycled CF-PC into industrial LFAM applications while highlighting the need for technological innovations to mitigate fiber degradation and enhance material performance for broader adoption. This critical step toward circular material practices in LFAM offers a pathway to reducing feedstock costs and environmental impact while maintaining functional performance in industrial applications.

additive manufacturing↗
DOE OSTI2023

Wave Energy Power Take-off Validation with a Hydraulicly Actuated Rotary Dynamometer and a Bi-Directional High-Power DC Supply

There are many organizations working towards the commercialization of wave energy converter technologies and are advancing their designs through the technology readiness levels (TRLs). A critical step before the field deployment of prototype wave energy converters is the validation of the subsystems and components that are contained in the wave energy converter through laboratory testing and performance characterization. In 2021, the National Renewable Energy Laboratory (NREL) developed and demonstrated a system for testing power takeoffs (PTO) with a low-speed, high-torque dynamometer and a grid-tied high-power DC power source and sink before field deployment. The hydraulic dynamometer allows for the simulation of PTO actuation from wave motion and is capable of a wide range of wave periods and heights which are represented as various speeds and torques from the dynamometer. The high-power bidirectional power supply allows for hardware in the loop and controller in the loop testing to be conducted on WEC power electronics. This presentation was made to describe the methods used by NREL research staff to test all components and sub-systems in the PTO of a novel wave energy converter before field deployment.

dynamometer↗
DOE OSTI2021

Quasi-periodic Oscillations in the TESS Light Curve of TX Col, a Diskless Intermediate Polar on the Precipice of Forming an Accretion Disk

One of the fundamental properties of an intermediate polar is the dynamical nature of the accretion flow as it encounters the white dwarf’s (WD’s) magnetosphere. Many works have presumed a dichotomy between disk-fed accretion, in which the WD accretes from a Keplerian disk, and stream-fed accretion, in which the matter stream from the donor star directly impacts the WD’s magnetosphere without forming a disk. However, there is also a third, poorly understood regime in which the accretion flow consists of a torus of diamagnetic blobs that encircles the WD. This mode of accretion is expected to exist at mass-transfer rates below those observed during disk-fed accretion, but above those observed during pure stream-fed accretion. We invoke the diamagnetic-blob regime to explain the exceptional Transiting Exoplanet Survey Satellite light curve of the intermediate polar TX Col, which transitioned into and out of states of enhanced accretion during Cycles 1 and 3. Power-spectral analysis reveals that the accretion was principally stream fed. However, when the mass-transfer rate spiked, large-amplitude quasi-periodic oscillations (QPOs) abruptly appeared and dominated the light curve for weeks. The QPOs have two striking properties: they appear in a stream-fed geometry at elevated accretion rates, and they occur preferentially within a well-defined range of frequencies (∼10–25 cycle day{sup −1}). We propose that during episodes of enhanced accretion, a torus of diamagnetic blobs forms near the binary’s circularization radius and that the QPOs are beats between the white dwarf’s spin frequency and unstable blob orbits within the WD’s magnetosphere. We discuss how such a torus could be a critical step in producing an accretion disk in a formerly diskless system.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Fabrication of 4H-SiC Low Gain Avalanche Detectors (LGADs)

Low gain avalanche detectors (LGADs) offer high temporal resolution for high energy particle detection, which is critical for next generation experiments in hadron colliders. While silicon LGADs (Si-LGADs) have rapidly matured in the last decade, research into silicon carbide (SiC) LGADs has only recently begun. By accounting for fundamental differences in material properties and fabrication processes, we present a prototype device design and process flow for 4H-SiC LGADs with etch-based isolation. Critical steps of the process flow and their results are discussed, including plasma etching, passivation, and the formation of low resistivity contacts. Electrical characterization (I-V, C-V) shows sufficient depletion of the device structure to demonstrate low-gain charge carrier multiplication.

High Energy Physics↗
DOE OSTI2021

Three-Dimensional CFD Investigation of Pre-Spark Heat Release in a Boosted SI Engine

Low-temperature heat release (LTHR) in spark-ignited internal combustion engines is a critical step toward the occurrence of auto-ignition, which can lead to an undesirable phenomenon known as engine knock. As such, correct predictions of LTHR are of utmost importance to improve the understanding of knock and enable techniques aimed at controlling it. While LTHR is typically obscured by the deflagration following the spark ignition, extremely late ignition timings can lead to LTHR occurrence prior to the spark, i.e., pre-spark heat release (PSHR). In this research, PSHR in a boosted direct-injection SI engine was numerically investigated using three-dimensional computational fluid dynamics (CFD). A hybrid approach was used, based on the G-equation model for representing the turbulent flame front and the multi-zone well-stirred reactor model for tracking the chemical reactions within the unburnt region. A recently developed best practice was also employed which keeps the well-stirred reactor model active throughout the entire simulation. This allowed for correct predictions of the previous cycle trapped residuals which have a considerable effect on the onset of LTHR. Multi-cycle simulations were conducted using Co-Optima alkylate and E30 fuels. The predicted in-cylinder pressure and heat release rate agreed well with the experimental data and served as validation for the CFD model. Following the initial validation, the dynamics of PSHR was discussed and a series of parameters of interest were assessed. First, the effect of exhaust valve temperature was investigated, qualitatively highlighting the importance of boundary conditions uncertainty. Further analyses were carried out on the effects of fuel properties, including laminar flame speed (LFS) and heat of vaporization (HOV). The results indicate that PSHR phasing is slightly advanced with lower LFS as more trapped unburnt fuel is made available for the next cycle. A similar trend was observed with lower HOV as less intense spray cooling led to higher mixture temperatures, i.e., higher mixture reactivity. Finally, a comparison of Co-Optima alkylate and E30 fuels was made using the pressure-temperature trajectory framework. It was shown that the differences between Co-Optima alkylate and E30 in PSHR tendency are correlated with both HOV and chemical effects.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2022

PRMI: A Dataset of Minirhizotron Images for Diverse Plant Root Study

Understanding a plant's root system architecture (RSA) is crucial for a variety of plant science problem domains including sustainability and climate adaptation. Minirhizotron (MR) technology is a widely-used approach for phenotyping RSA non-destructively by capturing root imagery over time. Precisely segmenting roots from the soil in MR imagery is a critical step in studying RSA features. In this paper, we introduce a large-scale dataset of plant root images captured by MR technology. In total, there are over 72K RGB root images across six different species including cotton, papaya, peanut, sesame, sunflower, and switchgrass in the dataset. The images span a variety of conditions including varied root age, root structures, soil types, and depths under the soil surface. All of the images have been annotated with weak image-level labels indicating whether each image contains roots or not. The image-level labels can be used to support weakly supervised learning in plant root segmentation tasks. In addition, 63K images have been manually annotated to generate pixel-level binary masks indicating whether each pixel corresponds to root or not. These pixel-level binary masks can be used as ground truth for supervised learning in semantic segmentation tasks. By introducing this dataset, we aim to facilitate the automatic segmentation of roots and the research of RSA with deep learning and other image analysis algorithms.

Xu, Weihuang↗