Selected publications

Compares four ways to interpret down-hole seismic measurements at three Thai sites, examining how overburden pressure affects inferred soil velocities.

Studies conformal prediction for quantifying uncertainty in HER2 status classification from breast-cancer immunohistochemistry images.

A VGG-based convolutional model predicts crash severity from Thai highway records for 2011–2023. Categorical features become grayscale images, with class balancing and Bayesian optimisation; evaluation uses F1-score and AUC.

Tests a simpler interpretation of explosion-generated seismic data against MASW at five Thai sites. Accuracy varies with changes in soil layers.

A camera-based traffic survey device detects, counts and classifies vehicles into 13 Department of Highways categories. Image processing and machine learning support automated traffic-volume collection.

Statistical association measures and a convolutional deep-learning model examine factors linked to fatalities in Thai highway crash records. The study highlights safety equipment, road-user type and lighting conditions.

Uses open spatial data and GIS network analysis to assess access to fire-response resources in Bangkok’s Phra Khanong district.

Examines truck overloading in Thailand through bridge weigh-in-motion data and enforcement evidence, linking infrastructure monitoring with transport policy.

A survey of 234 participants at Chulalongkorn University combines factor analysis and logistic regression to examine how safety awareness relates to recommending rental e-scooter services.

Tests GNSS receivers with inertial sensors at different pole tilts and levels of obstruction to evaluate positioning accuracy under NRTK surveying.

Applies GeoSAM to extract building footprints from 5 cm UAV true orthophotos and compares the results with manually delineated outlines at two Chulalongkorn campus areas.

Assesses GNSS rover horizontal positioning accuracy for network-RTK virtual-reference-station surveying, with relevance to cadastral measurements in Thailand.

Evaluates anaerobic fluidised bed reactors containing PVA media over 240 days, varying organic loading and hydraulic retention time without internal recirculation.

Surveys across five Silor routes in Bangkok compare travel patterns and service perceptions among users and non-users. Statistical classification identified four user groups with different needs and expectations.

Examines low-distortion map projections for provincial construction, focusing on balancing projection errors in urban areas.

Evaluates transfer learning with AlexNet and ResNet50 to locate and classify parasite eggs in images from a low-cost USB microscope.

Develops FracReconNet to reconstruct fractured proximal femurs in 3D from two radiographic views, using fracture representations and training-data augmentation.

A pavement-strength prediction study relates base materials, repeated axle loads, temperature and season to structural condition. Model outputs are compared with falling-weight deflectometer measurements from actual roads.

Compares onboard RTK, post-processed kinematic and precise-point-positioning workflows for UAV mapping using ground checkpoints.

HDM-4 simulations use Highway 344 data to examine pavement damage and maintenance costs over 20 years for four heavy-truck categories. The analysis supports exploration of maintenance-based truck charges.

Assesses how reference-station geometry and spacing relate to horizontal positioning accuracy across Thailand’s NRTK-VRS network.

Case studies of two urban rail systems and two tollways examine socio-economic benefits and pandemic-related financial risks in Thai public–private partnerships.

In a medical-imaging collaboration, introduces MANet: a U-Net-based network with residual learning and channel and spatial attention for liver-tumour segmentation in CT images.

Structural equation models compare how service quality, passenger confidence and health measures relate to bus-use intentions in Bangkok and Jakarta during the pandemic.

Structural equation modelling examined relationships between human resource practices, knowledge management and the performance of Thai infrastructure contractors. The analysis found positive links between these organisational factors in the firms studied.

Partial least-squares structural equation modelling examines how knowledge management and organisational learning mediate links between human-resource practices and sustainable performance in Thai construction firms.

Defines design principles and a software-assisted workflow for low-distortion map projections in construction projects, using PROJ parameters.

An online survey of 403 participants in Bangkok and nearby areas was analysed using structural equation modelling. The study examined how technology acceptance and environmental concern relate to battery-electric vehicle purchase intentions and use.

A structural equation model evaluates relationships among human-resource practices, knowledge management and non-financial performance in Thai infrastructure construction firms. Knowledge management is examined as a mediating factor.

Combines a Vision Transformer backbone with YOLOX and a feature-pyramid decoder to detect roadside assets in Thai highway images.

Repeated-load tests examined permanent deformation in recycled pavement materials treated with bitumen emulsion and cement. The laboratory results offer guidance for designing recycled pavement rehabilitation materials.

Examines the WGS-TM projection definition used for the Thai–Chinese high-speed railway and the parameters needed for consistent construction surveying.

Presents preliminary calculations for bolted and unbolted tunnel-lining joints, connecting joint flexural capacity with stiffness reduction and segment arrangement.

Combines a measured bridge influence line with iterative load estimation. Simulations and scaled experiments assess accuracy under road roughness and measurement noise.

Combines text segmentation with polygon offsets and border information to distinguish nearby and irregularly shaped text in natural images.

A modified fluid analogy model studies how station-area land use affects the balance of rail passenger flows. An MRT Blue Line case study explores mixed-use development as part of integrated transport and land-use planning.

Examines flow behaviour and treatment performance in a fluidised bed reactor, testing nitrate removal and anaerobic treatment of dilute wastewater.

Uses finite-element analyses to examine how uneven tail-void grout pressure changes tunnel-lining forces and deformation during construction.

Repeated-load triaxial tests investigate the effect of water content on crushed-limestone stiffness. The measured resilient-modulus behaviour informs mechanistic–empirical pavement design.

Laboratory research examines how crumb-rubber size and content affect asphalt deformation resistance when rubber replaces part of the aggregate through dry-process mixing.

Studies detection of the exposed cut ends of eucalyptus logs from images taken behind loaded timber trucks.

Ordered probabilistic models analyse 483 traveller surveys to study bus and informal-transport use in Phnom Penh. Costs, traveller characteristics, comfort and availability help explain differences in usage frequency.

Investigates image-based detection of missing fasteners for railway inspection.

Uses GPS observations to compare ionospheric electron-content variations during solar maximum and minimum at mid-latitude stations.

Two image-processing methods separate touching aggregate particles in asphalt cross-sections. Tests against a reference image found that size-based multiple erosion improved particle detection and size-distribution estimates.

A Green Growth Index combines the economic benefits and environmental performance of highway projects. The paper illustrates the assessment with a hypothetical project and a case study of Kanchanaphisek Road.

Field data from Bangkok’s Sathorn Road were used to model signal timing and reversible-lane operations in Synchro. The study compared proposed timings with observed police-operated signals to assess travel-time changes.

Combines colour, shape and line analysis to identify buildings or visible building parts in street-level images.

Evaluates observation duration and ionospheric conditions when using Department of Lands virtual-reference-station data to establish survey control.

Uses MiSeq sequencing to connect microbial communities with nitrate removal in a fluidized-bed reactor operated without internal recirculation.

Tracks shear-wave velocity in three sandy soils during drained triaxial compression, examining how stress conditions alter correlations with wave speed.

Uses a convolutional neural network to locate Thai text with different orientations in natural-scene images.

Investigates tuned mass dampers as a way to limit earthquake damage in buildings exposed to far-field ground motion.

An envelope method and Monte Carlo simulation model delay-risk insurance premiums that change with management responses during construction. Illustrative road and building projects demonstrate the approach.

Investigates Thai ionospheric-map corrections for improving network RTK GPS performance in cadastral surveying.

A stochastic revenue model and Monte Carlo option-pricing method adjust government guarantee thresholds over time. A BOT highway example explores allocation of revenue risk between government and concessionaire.

Presents prototype equipment for assessing mean pavement profile depth to support maintenance planning.

An analytic hierarchy process uses freight-forwarder surveys to compare seven Thai international ports. The indices cover infrastructure, port management, customs procedures and consistency of customs enforcement.

Compares methods for identifying individual mangrove trees and estimating their characteristics from airborne LiDAR, checked against field measurements in Thailand.

Combines laboratory and field evidence to estimate small-strain stiffness parameters for Bangkok clay, supporting calibrated geotechnical finite-element models.

Video observations on a Bangkok expressway identify three traffic states near an off-ramp queue. Changes in exit flow and lane changing explain shifts in bottleneck capacity and suggest traffic-responsive ramp control.

Tests a compact treatment process combining coagulation, flocculation, micro-bubble flotation and hydrocyclone separation using natural water from the river Marne.

Investigates semi-automatic faulting algorithms for highway engineering; the institutional bibliography verifies the conference work.

Compares oil-droplet tracking and ultrasound measurements with three-dimensional CFD simulations using k-epsilon and Reynolds-stress turbulence models in hydrocyclones.

Calibrates a finite-element representation of tunnel joints against full-scale segment tests, then examines how joint stiffness and segment count affect bending moments.

Compares constant and time-varying axle-load estimation in bridge weigh-in-motion simulations, testing sensitivity to speed, roughness and measurement error.

Combines measured train-induced strains with finite-element analysis to evaluate fatigue damage in long-serving Thai railway truss bridges.

Investigates a gravity-inspired image-processing step to separate noise from pavement cracks during segmentation.

Studies enhanced grid-cell analysis for detecting cracks in asphalt-surface images.

Develops a high-speed-camera method that tracks dyed oil droplets to measure water velocities inside conical and cylindrical hydrocyclones.

Validates the recovery of moving-vehicle axle loads from bridge bending moments through physical experiments and directly measured axle loads.

Observations at three freeway bottlenecks link rising vehicle density to subsequent capacity drops. A real-time density-estimation algorithm supports investigation of responsive traffic-control strategies.

Introduces an iterative method that separates static and dynamic axle loads in bridge response data, reducing sensitivity to the regularisation parameter in numerical examples.

Field observations and ramp-metering experiments examine how queues and lane changes reduce freeway merge capacity. Adjusting ramp inflow temporarily recovered higher discharge flows at the study site.

Develops a codebook model for separating foreground objects from changing backgrounds in video.

A boundary-element formulation models layered pavement and temperature-related stiffness gradients. Two Florida pavement examples examine how these gradients affect tensile stresses near surface cracks.

Tests hydrocyclones with and without a grit pot for runoff-water pretreatment, comparing solids separation and the volume and concentration of sludge.