Track Length Research Articles

Local Feature Performance Evaluation for Structure-From-Motion and Multi-View Stereo Using Simulated City-Scale Aerial Imagery

Ubiquitous low cost multi-rotor and fixed wing drones or unmanned aerial vehicles (UAVs) have accelerated the need for reliable, robust, and scalable Structure-from-Motion (SfM) and Multi-View Stereo (MVS) pipelines suitable for a variety of flightpath trajectories especially in degraded environments. Feature tracking being a core part of SfM and MVS, is essential for multiview scene modeling and perception, but difficult to evaluate in large scale datasets due to the lack of sufficient ground-truth. For large-scale aerial imagery, accurate camera orientation and dense 3D point cloud accuracy can be used to assess the impact of accurate feature localization and track length. We propose a novel view simulation (or synthesis) framework which generates visually realistic new unseen camera views for feature detection using known high fidelity camera poses for modeling. Seven state-of-the-art local handcrafted and learning-based features are quantitatively evaluated for robustness and matchability within the SfM and MVS pipelines using the open source COLMAP software. Our experimental results provide performance rankings of each feature, using twelve different evaluation metrics across three synthetic city-wide aerial image sequences. We show that recent learned features, SuperPoint and LF-Net, have not only reached the quality of the best handcrafted features like SIFT and SURF, but now outperform them in terms of more accurate 3D camera pose estimates and longer feature tracks. SuperPoint produces 1.51 meter average position error and 0.03° average angular error, while SIFT remains competitive (second best for pose and overall) with 1.78 meter and 0.11° errors respectively.

Disrupted Amygdala Connectivity Is Associated With Elevated Anxiety in Sensorineural Hearing Loss.

Background and Purpose: Hearing loss is associated with rising risks of emotional impairments, suggesting emotional processing networks might be involved in the neural plasticity after hearing loss. This study was conducted to explore how functional connectivity of the amygdala reconfigures in the auditory deprived brain and better understand the neural mechanisms underlying hearing loss-related emotional disturbances.Methods: In total, 38 chronic sensorineural hearing loss (SNHL) patients and 37 healthy controls were recruited for multimodal magnetic resonance imaging scanning and neuropsychological assessments. Voxel-wise functional connectivity (FC) maps of both the left and right amygdala were conducted and compared between the SNHL patients and healthy controls. The uncinate fasciculus (UF), an association fiber pathway, was reconstructed in both groups. The track number, mean track length, fractional anisotropy (FA) and mean diffusion values of the left and right UF were further quantified, respectively. Besides, Pearson's correlation analyses were conducted to investigate the relationship between the functional/structural abnormalities and the negative emotional states in SNHL patients.Results: The SNHL patients presented higher depressive and anxious levels compared to the healthy controls. Decreased FCs were detected between the amygdala and the auditory cortex, striatum, multimodal processing areas, and frontoparietal control areas in the SNHL patients. The amygdala was found to be structurally connected with several FC decreased regions through the UF. Moreover, the hypo-synchronization and the white matter impairment were both found to be associated with patients' elevated anxious status.Conclusions: These functional and structural findings depicted the reconfiguration of the amygdala in SNHL, which provided a new perspective toward the functional circuit mechanisms targeting the emotional impairments related to hearing loss.

CNN-based event classification of alpha-decay events in nuclear emulsion

Alpha-decay events in a nuclear emulsion are standard calibration sources for the relation between the track length and the kinetic energy in each emulsion sheet. We developed an efficient classifier that sorts such alpha-decay events from various vertex-like objects in an emulsion using a convolutional neural network (CNN). We trained the CNN using 15885 images of vertex-like objects, including 906 alpha-decay events, and tested it using a dataset of 46948 images including 255 alpha-decay events. The precision and recall scores of the classification using the previous method without a CNN for the same dataset were 0.081 ± 0.006 and 0.788 ± 0.056, respectively. In contrast, our trained models achieved an average precision score of 0.760 ± 0.006 for the test dataset, after extensively tuning the hyperparameters of the CNN. Moreover, for the model obtained, the discrimination threshold of the classification can be adjusted arbitrarily according to the trade-off between the precision and recall scores. Furthermore, the developed classifier obtained a precision of 0.571 ± 0.017 when the recall score was assigned a value of 0.788. Finally, the developed CNN method reduced the need for additional human visual inspection, required after classification, by a factor of approximately 1/7, compared to the former method without a CNN, proving the feasibility of the proposed classifier.

New Apatite Fission-Track Data from the Murmansk Craton, NE Fennoscandia: An Echo of Hidden Thermotectonic Events

For a long time, the thermal history of northeastern (NE) Fennoscandia in the Phanerozoic and Precambrian remained unknown, since no thermochronological studies were carried out within the Kola Peninsula area. Two years ago, we developed the first model of tectono-thermal evolution of the Kola Peninsula territory for the last 1.9 Gyr using a set of newly obtained apatite fission-track (AFT) and Ar/Ar thermochronological data. However, the low-temperature history of the most ancient tectonic unit of the northeastern part of the Kola Peninsula—the Archean Murmansk craton—remained poorly constrained due to the lack of AFT data. In this paper, we present the first results of AFT studies of 14 samples representing intrusive and metamorphic Precambrian rocks, located within the Murmansk craton of NE Fennoscandia. AFT ages and track length distributions indicate a similar tectono-thermal evolution of Precambrian tectonic units in NE Fennoscandia over the last 300 Myr. The AFT ages are distributed between ca. 177 and ca. 384 Ma; their median value, ~293 Ma, confirms the presence of a previously identified hidden thermal event that took place at about 300 Ma. However, a detailed analysis of the AFT age distribution shows the presence of three statistically distinguishable age components: 180–190 Ma (C1), 290–320 Ma (C2) and 422 Ma (C3). We assume that the relatively young AFT ages of C1 may originate from apatite crystals with low thermal resistivity. Remarkably, this value coincides with the initial stage of the Barents Sea magmatic province activity during large-scale plume-lithospheric interaction, as well as with the assumed age of an enigmatic remagnetization event throughout the Kola Peninsula. C2 ages can be observed in both the gabbroic and non-gabbroic samples, whereas C3 ages can only be found in gabbro. It is supposed that C2 ages, similarly to the Central Kola terrane, correspond to a cooling event related to the denudation of a thick sedimentary cover, representing a continuation of the Caledonian foreland basin towards NE Fennoscandia. C3 ages may be associated with a thermal event corresponding to the Caledonian collisional orogeny.

Change in risk of hospital admissions for ischemic heart disease after the implementation of a mass rapid transit system in Taipei

ABSTRACT Numerous epidemiologic studies demonstrated an association between an increase in levels of fine particles (particulate matter less than 2.5 um in diameter, PM2.5) and elevation in the number of hospital admissions for cardiovascular diseases. Air pollution levels including PM2.5 clearly decreased in Taipei City after the mass rapid transit (MRT) system began operations in 1996. The aim of this study was to investigate the extent of changes in the risk of daily hospital admissions for ischemic heart disease (IHD) over a 17-year period after the installation of a MRT system in Taipei. The full study was divided into Period 1 (1997–2000), total track length 65.1 km; Period 2 (2001–2008), total track length 75.8 km; and Period 3 (2009–2013), total track length 121.3 km. A time-stratified case-crossover analysis was conducted to estimate relative risk (RR) of hospital admissions for IHD for each 10 ug/m3 increase in PM2.5 for different periods. On cool days, the associated RR of IHD for Period 3 was consistently lower compared to period 2 in both our single- and two-pollutant models. However, the daily risk for IHD admissions was found to be significantly higher for period 3 compared to period 2 in our single-pollutant model and in our two-pollutant models (PM2.5+ SO2) on warm days. The basis for this difference is unknown. Data suggests that an MRT system may provide substantial health benefits, a finding that may be helpful to urban communities, urban planners, and public health specialists.

Path Planning Algorithm of Dijkstra-Based Intelligent Aircraft under Multiple Constraints

Aiming at the rapid planning of the optimal flight path of the intelligent aircraft, considering the error constraints and correction probability constraints, a model for intelligent aircraft path planning under multiple constraints is constructed, and a global search algorithm based on Dijkstra algorithm is proposed to solve the model. By calculating the residual error and restricts flight distance, the basic Dijkstra algorithm is improved to make it more adaptable to solve the path planning under multiple constraints. At the same time, simulation experiment is conducted with the optimal goal of the shortest track length and satisfying the error constraints. The experimental results show that the aircraft passed a total of 18 correction points when it reached the destination. The total track length was 144 287.932 m, the vertical position error was 17.254 units, and the horizontal position error was 6.420 units. The results meet the error requirements. The results show that the intelligent aircraft path planning model and Dijkstra-based global search algorithm with multiple constraints are reasonable in solving such problems.

Neuronal trafficking dynamics and regulators

AbstractBackgroundNeurons are highly polarized cells continuously relying on intracellular transport, through which they communicate with the external world. Among the somatodendritic and axonal compartments many proteins with different functions are sorted to reach long distances. The relevance of neuronal transport in ensuring cell development and homeostasis maintenance raises a big interest on this topic. Moreover, several neurodegenerative diseases, such as Alzheimer’s or Amyotrophic Lateral Sclerosis, exhibit significant axonal pathology, where transport impairment seems to have a crucial influence. However, the role of this process in neurodegeneration, remains poorly understood. In this study we aim to characterize neuronal transport dynamics of specific proteins, and the molecular machinery regulating this process. Furthermore, we aim to explore transport from the pathological point of view, evaluating the effect that mutant proteins (e.g. TDP43‐related) may have on it.MethodHuman Neural Stem Cells derived neurons were differentiated for 30 DIV. Plasmids encoding for different GFP‐coupled proteins were designed to express those who are involved in main neuronal functions (e.g. synaptic activity) and transfected in neurons. Time‐lapse movies of cargoes were acquired to detect changes in movement. Data analysis was performed using object segmentation and tracking algorithms. Raw data were processed to describe the main motion parameters: average velocity, directionality, track length, and others. Post‐imaging Immunocytochemistry was performed to assess cargoes localization and discriminate between axonal and dendritic transport.ResultAmyloid Precursor Protein (APP) cargoes with higher velocity in anterograde direction were observed compared with the retrograde one; meanwhile for Synaptophysin, a major population of stationary particles was found. Moreover, sorted data for axonal and dendritic transport description shown faster APP particles in axonal anterograde movement. Relying on the movement dynamics description, we decided to screen for protein‐protein interactions, using GFP‐trap and CoIP approaches followed by Mass Spectrometry analysis, which can shed light on possible partners involved in the transport regulation.ConclusionOur work describes an in vitro model for neuronal transport study and highlights the different behavior of proteins with various physiological functions. A complete view will elucidate details about transport regulation and will open a way to understand its role in neuropathological processes.

The HLTF\u2013PARP1 interaction in the progression and stability of damaged replication forks caused by methyl methanesulfonate

Human HLTF participates in the lesion-bypass mechanism through the fork reversal structure, known as template switching of post-replication repair. However, the mechanism by which HLTF promotes the replication progression and fork stability of damaged forks remains unclear. Here, we identify a novel protein–protein interaction between HLTF and PARP1. The depletion of HLTF and PARP1 increases chromosome breaks, further reduces the length of replication tracks, and concomitantly increases the number of stalled forks after methyl methanesulfonate treatment according to a DNA fiber analysis. The progression of replication also depends on BARD1 in the presence of MMS treatment. By combining 5-ethynyl-2′-deoxyuridine with a proximity ligation assay, we revealed that the HLTF, PARP1, and BRCA1/BARD1/RAD51 proteins were initially recruited to damaged forks. However, prolonged stalling of damaged forks results in fork collapse. HLTF and PCNA dissociate from the collapsed forks, with increased accumulation of PARP1 and BRCA1/BARD1/RAD51 at the collapsed forks. Our results reveal that HLTF together with PARP1 and BARD1 participates in the stabilization of damaged forks, and the PARP1–BARD1 interaction is further involved in the repair of collapse forks.

Wide-range tracking and LET-spectra of energetic light and heavy charged particles

We developed a highly-selective technique to measure the energy loss and linear-energy-transfer (LET) spectra of energetic charged particles in high-resolution and over a large collection of particle-event types. Precise and wide-range spectral and tracking measurements were performed with a single semiconductor pixel detector. The quantum-counting sensitivity, high-granularity and per-pixel spectrometric response of the Timepix ASIC chip enable the detailed spectral-tracking registration of single charged particles across the detector semiconductor sensor. Both the deposited energy along the particle trajectory (energy loss) and the path length of the particle track across the semiconductor sensor are precisely measured for each particle. This allows for the determination of the particle LET in silicon in high accuracy and over a wide-range of energies, particle types and directions. The tracking and energy loss response together with the resolving power at the particle-event level make it possible to selectively provide LET distributions of the light and heavy charged particle components in mixed-radiation and omnidirectional fields. This technique applies to energetic (E > 10 MeV/u) charged particles generating tracks greater than the pixel size and incident at non-perpendicular direction (>20∘) to the sensor plane. The technique applies also to electrons of energy above few MeV as well as highly energetic and minimum-ionizing-particles (MIPs). We make use of existing and in part newly collected data at well-defined radiation fields with proton and light ion beam accelerators. Flexible measurements, ease of deployment and online response are possible by the use of compact readout electronics such as the miniaturized radiation camera MiniPix (size < 8 cm, weight < 50 g) operable by any PC. Results are given for protons and light ions (He, C) of selected energies above 10 MeV/u and directions (2π FoV). We include also electrons (20 MeV). Selective and detailed LET spectra are produced over a wide range (10−1 to 102 keV/μm) in silicon.

Calibration of FHCal with cosmic muons at the BM@N experiment

The forward hadron calorimeter (FHCal) is one of the sub-detectors of the BM@N experimental setup at JINR, Dubna. It consists of 54 lead-scintillator “sandwich” type modules of two types with the transverse sizes 20 × 20 cm 2 and 15 × 15 cm 2. These two types of modules are subdivided into 10 and 7 individual longitudinal sections, respectively. Each section provides the independent light and amplitude signal readout with one silicon photomultiplier (MPPC). High signal to noise ratio of MPPC allows to detect cosmic rays with low energy depositions in FHCal longitudinal sections. A method for cosmic muon track reconstruction is discussed. A procedure for energy calibration based on muon track length and energy deposition in each section is proposed. Experimental results of FHCal cosmic calibration are presented.

Fatigue in the Indonesian rail industry: A study examining passenger train drivers

ObjectiveThe railway industry in Indonesia has grown substantially in the past five years, with greater network and tracks length that will double in 2030. There has been a growing concern, however, that fatigue among front-line personnel have not been addressed satisfactorily. This study aimed at evaluating the degree of fatigue among passenger train drivers and suggested practical recommendations to the Indonesian Railway Company. MethodA total of 32 train drivers (driving for about 4 h during the morning or afternoon shifts) agreed to participate in the study. Subjective and objective measures of fatigue (and sleepiness) were obtained prior to, during, and immediately following the end of a duty. Fatigue and driving performance were also determined based on (off-line) analysis of continuous video recordings of the drivers performing the duty. ResultsFatigue generally increased as a function of time and, for some, 2 h of driving was sufficient in inducing substantial level of fatigue increase. Unlike the morning shift, the afternoon shift was characterized with increased undue fatigue and unacceptable level of performance decrements. ConclusionIn general, fatigue was not found among train drivers working the morning shift. For these drivers, however, there were a couple of individuals with excessive levels of fatigue. The effects of fatigue were more apparent for the afternoon shift. This is possibly due to the degraded sleep quality as a result of split rest periods obtained prior to the duty. It is suggested in this study that the company pay a much closer attention to the work-sleep schedules, and to improve sleep quality by redesigning better sleeping arrangements.

Software for designing a port railway station

The subject of the research is programming the problem of designing a port station and calculating its key parameters. The relevance of the research topic is related to the need to automate a complex of complex calculations that can be performed more quickly and accurately using a software product. Methods dynamic and linear programming, syntax and tools of the Visual Basic 6.0 programming language, analytical, marketing, and project methods, materials of general access and official sites, own research used. The software product for automating the calculation of technical and operational parameters of the port station is described. The results of the calculation can be applied when constructing a continuous schedule for the operation of a transport hub, a diagram of daily wagon flows in the station. Using the program, you can determine the number and length of tracks for receiving, sending and sorting trains by station. The software product can be useful both in the educational process of secondary and higher education, and in practical business when making business decisions on the organization of the work of both the port station and the transport hub as a whole.

South Atlantic passive margin evolution: A thermochronology case study from the Rio de Janeiro-Três Rios section, SE Brazil

The southeastern Brazilian passive margin records a complex post-rift evolution, with two parallel high-elevation features formed after the opening of the South Atlantic. We applied apatite fission track (AFT) and U–Th/He (AHe) low temperature thermochronology to constrain the thermo-tectonic history of the Serra do Mar escarpment in the area of Rio de Janeiro state. New AFT central ages for basement areas collected from a N–S transect orthogonal to the shoreline between the cities of Rio de Janeiro and Três Rios, range between 98.5 ± 5.3 and 54.1 ± 4.2 Ma, with mean track lengths between 12.34 ± 0.40 and 14.63 ± 0.17 μm. Uncorrected AHe ages lie between 68.1 ± 5.9 and 60.2 ± 7.3 Ma and are consistent with AFT results. Inverse thermal history models constrained by AFT and AHe data imply the earliest cooling onset from the Barremian (Early Cretaceous), with steady rates more common for samples closer to coastal areas. Maximum depths of denudation are between 2.5 and 4.5 km. Published thermochronological data from adjacent areas combined with the new results show a seemingly simpler post-rift evolution for the area, although suggesting structural control of age distribution and exhumation.

Range-of-motion affects cartilage fluid load support: functional implications for prolonged inactivity

Joint movements sustain cartilage fluid load support (FLS) through a combination of contact migration and periodic bath exposure. Although there have been suggestions that small involuntary movements may disrupt load-induced exudation during prolonged inactivity, theoretical studies have shown otherwise. This work used well-controlled explant measurements to experimentally test an existing hypothesis that the range-of-motion must exceed the contact length to sustain non-zero FLS. Smooth glass spheres (1.2-3.2mm radius) were slid at 1.5mm/s (Péclet number >100) against bovine osteochondral explants under varying normal loads (0.05-0.1N) and migration lengths (0.05-7mm) using a custom instrument. In situ deformation measurements were used to quantify FLS. Non-zero FLS was maintained at migration lengths as small as 0.05mm or <10% the typical contact diameter. FLS peaked when track lengths exceeded 10 times the contact diameter. For migration lengths below this threshold, FLS decreased with increased contact stress. Migration lengths far smaller than the contact diameter can sustain non-zero FLS, which, from a clinical perspective, indicates that fidgeting and drifting can mitigate exudation and loss of FLS during prolonged sitting and standing. Nonetheless, FLS decreased monotonically with decreased migration length when migration lengths were less than 10 times the contact diameter. The results demonstrate: (1) potential biomechanical benefits from small movement (e.g., drifting and fidgeting); (2) the quantitative limits of those benefits; (3) and how loads, movement patterns, and mobility likely impact long term FLS.

Field trials of a methodology for locomotive brake testing to assess friction enhancement in the wheel/rail interface using a representative leaf layer

There are two options for testing the performance of rail vehicles and the application of material to the wheel/rail interface to influence friction, and therefore braking performance, in a low adhesion environment: laboratory work which is not representative of conditions in the field, and expensive track tests using fully instrumented modern vehicles largely using low adhesion simulants that are not representative of those that occur in the real world and usually applied over great lengths of track. This paper presents a third option: low-cost track tests. We determined the minimum of equipment and resources needed to produce low adhesion using a method more representative on Autumn conditions on rail and monitor the performance of rail vehicles when braking on it. This was mainly targeted at use for testing products applied to the wheel/rail interface to enhance traction, hence the need for real low adhesion causes. We present the methodology that we have developed, and some initial results showing that using it we can generate a low adhesion scenario when braking from 10 mph. This can be used to develop mitigation solutions for low adhesion in a more cost effective manner.

A method to measure the integral vertical intensity and angular distribution of atmospheric muons with a stationary plastic scintillator bar detector

We present a method to measure the integral vertical intensity and angular distribution of atmospheric muons using a stationary thick plastic scintillator bar detector with PMTs at both of its ends over exposures of a few hours in a laboratory setting. The total flux is obtained from the event rate and the angular distribution is inferred from the distribution of the recorded pulse charges, which are correlated with the track length inside the plastic scintillator. A muon generator algorithm was developed and used together with a simple simulation including the effects of the detector resolution, nonlinear saturation of the PMTs, and laboratory building coverage. As a proof of principle we made a measurement assuming a standard cosnθ omnienergetic angular distribution with azimuthal isotropy and different models for the energy spectrum of the source. Using measurements at 5 different azimuthal orientations, we measure an average exponent n=(1.90±0.06(stat)±0.10(syst)), and an average vertical intensity of I0=(101.2±1.8(stat)±5.5(syst))m−2s−1sr−1 at the location with geographic coordinates 19.33°N 99.19°W, altitude 2268 m above sea level, and geomagnetic cut-off of 8.2 GV.

Verification and Validation of the Geant4 Monte Carlo Code Toolkit for DEMO TBR Evaluations

Verification and Validation analyses of the Monte Carlo code toolkit Geant4 were conducted with nuclear data from the JEFF-3.3 library to show the suitability for fusion applications. We present benchmarking against data of the engineering design oriented HCPB (Helium Cooled Pebble Bed) Tritium Breeder Module (TBM) mock-up experiment taken from the SINBAD compilation and nuclear analyses in comparison with MCNP of the DEMO reactor equipped with a HCPB breeding blanket. Missing functionalities for Geant4 were developed such as a track length detector with multiple energy bins and tally multiplication, reflective boundaries, and statistical checks in a post-processing step. The activation foil measurements in the TBM mock-up revealed Geant4 underestimates the uncollided and fast neutron flux increasingly with penetration depth, while being consistent with MCNP for the low energy neutron flux. Geant4 underestimates the experimental tritium activity in the breeder pellets by up to 2.7% more than MCNP, which is within the experimental uncertainty. The DEMO analyses yielded a 0.97% lower Tritium Breeding Ratio (TBR) of Geant4 than MCNP for a homogenized breeder blanket, and a 1.46% higher TBR of Geant4 than MCNP for a mostly heterogenic blanket. This good agreement with MCNP demonstrates Geant4’s suitability for DEMO TBR evaluations.

Electric resistance as a sensitive measure for detecting graphene wear during macroscale tribological tests

Super-high mechanical strength and excellent lubrication property at the nanoscale render graphene a promising ultra-thin solid lubricant. However, macroscale friction tests have shown that wear of graphene often occurs resulting in failure of interface lubrication. As graphene coatings are atomically thin, a sensitive and reliable detection method for monitoring wear has been a great challenge to date. In this work, we performed tribological experiments on monolayer graphene on SiO2/Si while simultaneously measuring the in-plane electric resistance of graphene during the sliding tests. Our experiments showed that the electric resistance exhibited a sudden increase signifying the onset of graphene wear well before the macroscale friction increased rapidly. The variation of in-plane electric resistance was found to depend on the wear track length and relative orientation. Longer wear track and wear track perpendicular to the current would typically lead to higher increase in graphene electric resistance. Our work demonstrates that monitoring in-plane electric resistance can be a simple, sensitive and predictive method for detecting graphene wear, which makes it a useful inspecting technique for a wide range of mechanical applications.

Novel borosilicate glass system: Na2B4O7-SiO2-MnO2: Synthesis, average electronics polarizability, optical basicity, and gamma-ray shielding features

A novel glass system with a chemical composition described by the formula 50SiO2– (48+x) Na2B4O7 – x MnO2 (0≤x≤2 mol %) was fabricated using a melt-quenching method. An XRD diffractometer procedure was applied to check the state of these glass samples. Obtaining the physical and optical values of the manganese borosilicate glass system can help develop solid-state devices and optical memory equipment. The energy gap shifting to higher energies may be due to the formation of bridging oxygens (BO). Urbach energies decrease in the range between 0.736 and 0.485 eV, with increasing the borax concentration between 48-50 mol %, respectively. The polarizability and optical basicity were investigated based on the experimental values of the refractive index for the fabricated glass samples. The polarizability and basicity were found to have the same trend as the refractive index. Monte Carlo N-particle transport code (MCNP-5) was applied to evaluate the gamma-shielding features for the investigated glass samples at gamma-photon energies between 0.248 and 1.406 MeV. The linear attenuation coefficient (LAC) was simulated based on the average track length of the gamma photons inside the fabricated glass samples. The highest LAC was achieved for the G5 glass sample, which has no MnO2 and changed between 0.330-0.153 cm−1 for photons with energy ranged between 0.248 and 1.406 MeV, respectively.

Elucidating the effect of preheating temperature on melt pool morphology variation in Inconel 718 laser powder bed fusion via simulation and experiment

In laser powder bed fusion (L-PBF) additive manufacturing, the mechanical performance, microstructure and defects of fabricated parts are closely associated with the melt pool morphology, e.g., its dimension and shape through the building process. Past studies have largely focused on how the process parameters such as laser power and scan speed affect melt pool characteristics. In this study, the melt pool morphology variation as a function of preheating temperature in the conduction, transition, and keyhole regimes and the underlying mechanisms in each regime are investigated through ex-situ sample characterization and computation thermal fluid dynamics (CtFD) simulation. Single tracks with different combinations of laser power and scan speed are deposited on an Inconel 718 bare plate preheated to a temperature range of 100–500 °C in the experiment. Significant changes are observed in melt pool morphology as a function of preheating temperature from optical measurements of melt track cross sections. The depth of melt pool in the three regimes increases monotonically with preheating temperature, e.g., at 500 °C, the experimental melt pool depth is increased by 49% in conduction regime, 34% in transition regime and 33% in keyhole regime, respectively, while the variation of melt pool width in each regime does not all follow an increasing trend but depends on the melt pool regimes. Melt pool width variation in the conduction and transition regimes is found to depend on the enhanced heat conduction directly related to temperature dependent thermal properties. Through validated CtFD simulations, it is found that in the keyhole regime the evaporation mass, recoil pressure, and laser drilling effect is enhanced with higher preheating temperature, which gives rise to a deeper melt pool. The simulations also reveal that preheating temperature significantly elongates the melt track length due to the increased flow rate and strong recoil pressure that accelerates the backward flow.