Accurate, reliable, and geological stratification and associated soil property estimation is critical for the safe and cost-effective design of geotechnical infrastructure. However, the inherent spatial heterogeneity of geomaterials, coupled with the scarcity, incompleteness of site investigation data, poses significant challenges for conventional site characterization methods. This study presents an integrated probabilistic framework that integrates Gaussian Mixture Models (GMM), Markov Random Fields (MRF), and a Hierarchical Bayesian Model (HBM) to address these challenges. The proposed framework first applies GMM to classify borehole records into probabilistic geological clusters in the feature space, followed by MRF-based three-dimensional stratigraphic modeling to incorporate spatial continuity and geological prior knowledge. For each cluster, soil parameters are inferred at unobserved locations using an HBM calibrated with relevant Big Indirect Data (BID), and the final predictions are obtained through probability-weighted aggregation across clusters. The framework is validated using two benchmark problems from the Tokyo Airport soft soil dataset: (i) reconstruction of undrained shear strength profiles from partial observations, and (ii) estimation of missing mechanical parameters under incomplete-testing scenarios. Comparative evaluations against existing benchmark approaches show that the proposed GMM–MRF–HBM framework captures depth-dependent variability more effectively, achieves lower prediction errors. These findings highlight the potential of the framework as a robust and generalizable tool for data-driven site characterization in sparse and heterogeneous geotechnical settings.

Rapid urbanization has caused numerous waste soil landfills. While most stratigraphic modeling has focused on natural geological formations, the stratigraphic characteristics of man-made landfills remain underexplored, posing potential safety risks. This study aims to characterize the stratigraphic heterogeneity and model the stratigraphic configuration of waste soils using multiple unmanned aerial vehicle (UAV) data. The reverse stockpiling method, combined with a progressive front dumping strategy, forms inclined dumping surfaces and produces rotated anisotropic strata in landfills. A novel UAV-based stratigraphic modeling method is developed to encode these anisotropic features and perform the stratigraphic interpolation by integrating Markov random fields and Bayesian approaches. UAV imagery is used to interpret the spatial distribution of visible soils. Three anisotropic potential functions are custom-designed to reflect the spatial constraint structure of inclined strata. Bayesian model comparison approach identifies model parameters, i.e., spatial correlation lengths, and selects the most plausible potential functions and stratigraphic profiles. The method is validated through a real-world landfill case. Results show the largest spatial correlation length along the strike direction of the inclined surface, followed by the dip direction, and the smallest along the normal direction, reflecting the level of anisotropy. Simulated stratigraphic profiles align with the observed inclined stratum structure in fields. This study provides a new approach and a good dataset for modeling the stratigraphic heterogeneity of waste soil, contributing to the safety assessment of man-made landfills.

Abstract Buried pipelines face threats from electrochemical corrosion, dynamic geological, and geo-hydrological hazards. Current fragmented datasets and associated standards hinder streamlined and comprehensive one-stop risk assessment. This paper proposes a dynamic database framework that integrates pipeline physical attributes with electrochemical metrics (e.g., close-interval potential surveys, DC voltage gradients, in-line inspection data), geohazard indicators (earthquake events, landslide susceptibility indices, shear-wave velocity profiles), and hydrological factors (river proximity, channel-scour metrics). Automated data harvesters leverage FDSN web services, United States Geological Survey (USGS) APIs, and Google Earth Engine to continuously ingest and normalize new datasets when available. This framework eliminates fragmentation across traditionally isolated data sources by providing a unified, spatiotemporally indexed platform that supports complex queries for corrosion, geological, and hydrological related pipeline risk modeling and visualization. The primary outputs is a list of geo-referenced objects, where each pipeline segment is represented as a spatial feature enriched with multi-factor risk attributes. This enables rapid visualization in QGIS, ArcGIS, or web libraries like Leaflet, and supports advanced spatial operations such as hotspot detection and proximity-based alerts. The standardized format allows integration with rule-based systems and machine learning processing pipelines. Ultimately, this extensible, scalable framework lays the foundation for pipeline digital twin applications and data-driven pipeline integrity management.

Abstract Detecting corrosion concealed beneath protective coatings remains a significant challenge for nondestructive evaluation of assets. In this study, The authors introduce a thermal imaging–based approach that applies transient heating to coated steel specimens and analyzes the resulting surface response using pixel-wise curve fitting and statistical clustering. A thermal infrared camera captured thermal image sequences during controlled halogen lamp excitation. Each pixel’s temperature–time series was fit with a sigmoid model, producing parameter maps that quantified amplitude, slope, and inflection time of the thermal response. Rust-affected areas exhibited delayed but amplified heating response profiles, whereas intact coated regions displayed more moderate responses. Gaussian Mixture Model (GMM) clustering of the fitted parameters successfully separated rust from non-rust regions, providing automated detection and classification without prior labeling. The results demonstrate that transient thermography, combined with sigmoid fitting and clustering, can effectively reveal corrosion hidden beneath coatings, offering a nondestructive, non-contact, and uncertainty-aware tool for integrity assessment of coated metallic surfaces.

This study proposes a fully coupled conditional simulation framework for jointly characterizing geological uncertainty and geotechnical variability under sparse site investigation data. In conventional practice, soil-category simulation (Task 1, T1) and soil-property simulation (Task 2, T2) are treated in a decoupled manner, conditioning on observed categorical data (L) and continuous soil property data (X) separately. The proposed framework departs from this paradigm by adopting a fully coupled strategy in which both L and X are simulated by conditioning jointly on {L, X}, thereby explicitly accounting for their statistical dependence. Implementing such a framework requires knowledge of site-specific X-L and X-X correlations, which are often weakly identifiable from sparse target-site data. To address this challenge, a modified hierarchical Bayesian model (HBM) is developed to learn these correlation characteristics from a newly compiled global soil database and transfer them to the target site as an informative prior. The framework is further equipped with an efficient conditional simulation algorithm for X, enabling practical three-dimensional applications. The performance and advantages of the proposed framework are demonstrated through a real case study and comparative analyses with existing methods.

Purpose This paper aims to contribute primarily in two areas: using multiple new strategies to devise an improved sand cat swarm optimization (ISCSO) algorithm with superior performance and exploring its applicability to the path planning issue that requires finding a safe route with the shortest length for an agricultural robot. Design/methodology/approach This paper designs and introduces multiple new strategies to modify the sand cat swarm optimization (SCSO) algorithm from different perspectives. Subsequently, 23 well-known standard benchmark function experiments and CEC2021 function experiments are performed using the ISCSO algorithm and another five approaches, encompassing the SCSO algorithm, the Harris Hawks optimization (HHO) algorithm, the GWO, the Snake Optimizer (SO) and the Zebra Optimization Algorithm (ZOA). Then, the results are analyzed to showcase the efficacy and superiority of the ISCSO algorithm. On this basis, we also explore the effect of applying the ISCSO algorithm to puzzle out the agricultural robot path planning issue. Findings All experimental results manifest that, except for a few functions among the 23 standard benchmark function experiments and CEC2021 function experiments, the ISCSO algorithm performs better overall than the other five algorithms with regard to optimization ability, convergence rate and stability. Moreover, the ISCSO algorithm is better suited for addressing the path planning issue encountered by the agricultural robot and exhibits stronger optimization ability in comparison to the SCSO algorithm. Originality/value This paper devised a novel improved SCSO algorithm with better performance and explored its applicability to the path planning issue that requires finding a safe route with the shortest length for an agricultural robot.

The integrity of underground pipelines is vital for the safe transport of resources like oil, gas, and water, but they face a variety of risks such as geohazard, third-party damage, and corrosion. These risks, coupled with the unpredictable nature of geohazards, demand innovative management strategies. This paper reviews existing pipeline risk management frameworks, which often rely on static data sets and reactive approaches. The review explores methodologies for pipeline hazard identification, risk assessment models, data integration techniques, and decision support tools. The paper also evaluates publicly available data sources, including those from NOAA, USGS, and NASA, alongside industry-specific databases and technologies like GIS-based visualization and machine learning. By identifying gaps and opportunities in current practices, this study establishes the current understanding for developing a dynamic, data-driven system to enhance proactive and predictive pipeline risk management. The findings offer a straightforward approach to enhancing pipeline safety, reliability, and resilience in challenging environments and operations.

This study proposes a Bayesian probabilistic method for calibrating digital fringe projection systems, addressing both aleatoric and epistemic uncertainties that can impact measurement accuracy and precision. The approach centers on Bayesian inference, which is particularly important for achieving quantified uncertainty in high-dimensional parameter space. In the context of 3D reconstruction, accuracy and precision are critically important because uncertainties can propagate and accumulate, possibly leading to significant random and systematic errors in final results. These errors can compromise the overall usability of the reconstruction results. Through comparative analysis with traditional maximum-likelihood-based calibration approach, for the first time to the best of the authors’ knowledge, our study demonstrates that the Bayesian approach not only enhances the confidence on uncertainty quantification but also significantly improves the estimated precision of 3D coordinate measurements. These improvements are particularly important in applications such as quality control, vision-based welding systems, and other precision-dependent tasks, where reliable and accurate 3D measurement are essential. The findings underscore the superiority of Bayesian inference in 3D measurement applications, making it a more dependable choice for producing high-accuracy and high-precision reconstructions.

In sensor metrology, noise parameters governing the stochastic nature of photon detectors play critical role in characterizing the aleatoric uncertainty of computational imaging systems such as indirect time-of-flight cameras, structured light imaging, and division-of-time polarimetric imaging. Standard calibration procedures exists for extracting the noise parameters using calibration targets, but they are inconvenient or impractical for frequent updates. To keep up with noise parameters that are dynamically affected by sensor settings (e.g. exposure and gain) as well as environmental factors (e.g. temperature), we propose an In-Scene Calibration of Poisson Noise Parameters (ISC-PNP) method that does not require calibration targets. The main challenge lies in the heteroskedastic nature of the noise and the confounding influence of scene content. To address this, our method leverages global joint statistics of Poisson sensor data, which can be interpreted as a binomial random variable. We experimentally confirm that the noise parameters extracted by the proposed ISC-PNP and the standard calibration procedure are well-matched.

Corrosion pit nucleation and propagation in underground pipelines vary by geographic position along the pipeline right of way (RoW). Effective pipeline integrity management requires both the modeling of corrosion initiation/propagation and the integration of field-acquired data. To mitigate external corrosion, onshore pipelines use barrier coatings and either sacrificial or impressed current cathodic protection (CP) systems. Pipeline integrity is monitored through direct and/or indirect assessments. In this work, we focused on leveraging indirect methods, such as close interval potential surveys (CIPS) and direct current voltage gradient (DCVG). CIPS evaluates CP effectiveness by identifying deficiencies in CP-equipped structures via pipe to soil potential and detecting potential 'hot spots' in those with insufficient CP. Meanwhile, the DCVG method identifies coating flaws by measuring voltage gradients using reference electrodes placed in the soil. But the reliability of defect detection is a practical concern. To enhance corrosion assessment reliability, we propose a Bayesian framework combining indirect inspection data, corrosion science, and artificial intelligence. Using Stochastic Variational Inference, the framework predicts corrosion severity and rates. Validation against inline inspection data demonstrates its accuracy and confidence in identifying corrosion-prone areas, thus improving pipeline integrity management and reducing failure risks.

Site response is of great concern in the central and eastern United States, particularly in the New Madrid Seismic Zone. Strong underlying impedance contrasts give rise to site resonances, the characterizing of which has been a major focus of earthquake-related research efforts in the New Madrid region for nearly three decades. To account for site responses and their spatial variability, this research utilized Gaussian Process to produce a series of 3D random fields for subsurface characterization, based on Vs profiles from seismic reflections and refractions, mapped geologic units, and digital elevation model data sets. The results of the random field realizations were used to calculate and map the two primary site resonance parameters, the fundamental site period fT and the amplification at that period, 0A. The results present not only the mean estimation of ground-motion site responses, but also their spatial variability as explicitly depicted on the mapping products.

Monitoring and preventing Geohazard such as landslides is a critical aspect of infrastructure engineering, essential for safeguarding infrastructure and communities. Effective slope stability analysis and failure monitoring demand the integration of comprehensive, field-acquired data to capture the complex dynamics of soil-structure interaction. While traditional methods such as remote sensing and aerial monitoring are commonly employed for geohazard assessment, they often fall short in providing the real-time capabilities necessary for timely intervention for risk assessment. Our approach leverages a network of multiple cameras to generate dynamic 3D point clouds, enabling continuous and detailed monitoring of soil stability and topographical changes. By producing scale-invariant 3D point cloud models for each frame, our method ensures the consistency and accuracy of geometry data, effectively eliminating physical scale variances that can occur during point cloud registration. This advanced system not only delivers accuracy on par with LiDAR but also provides the significant advantage of real-time analysis, making it possible to detect potential issues as they develop. By integrating this cutting-edge technology, we can significantly improve our capacity to forecast potential failure risks, allowing for more proactive management and mitigation of geohazards.

Quantifying stratigraphic uncertainty is crucial for reliable risk assessment and informed decision-making in geotechnical and geological engineering. However, accurately modeling complex stratigraphy—especially in heterogeneous settings influenced by irregular deposition—remains a challenge, particularly with limited site data. This study introduces a novel solution, modeling stratigraphy as a categorical random field and using image warping to transform non-stationary random fields into stationary ones, facilitating fast and realistic stochastic simulation. The method demonstrates high accuracy and computational efficiency in capturing complex stratigraphic profiles with quantified uncertainty. Validation through synthetic and real-world cases confirms the approach’s reliability and applicability.

Automatic Heuristic Design (AHD) is an effective1 framework for solving complex optimization prob-2 lems. The development of large language mod-3 els (LLMs) enables the automated generation of4 heuristics. Existing LLM-based evolutionary meth-5 ods rely on population strategies and are prone6 to local optima. Integrating LLMs with Monte7 Carlo Tree Search (MCTS) improves the trade-off8 between exploration and exploitation, but multi-9 round cognitive integration remains limited and10 search diversity is constrained. To overcome these11 limitations, this paper proposes a novel cognitive-12 guided MCTS framework (CogMCTS). CogMCTS13 tightly integrates the cognitive guidance mecha-14 nism of LLMs with MCTS to achieve efficient au-15 tomated heuristic optimization. The framework16 employs multi-round cognitive feedback to incor-17 porate historical experience, node information, and18 negative outcomes, dynamically improving heuris-19 tic generation. Dual-track node expansion com-20 bined with elite heuristic management balances the21 exploration of diverse heuristics and the exploita-22 tion of high-quality experience. In addition, strate-23 gic mutation modifies the heuristic forms and pa-24 rameters to further enhance the diversity of the so-25 lution and the overall optimization performance.26 The experimental results indicate that CogMCTS27 outperforms existing LLM-based AHD methods in28 stability, efficiency, and solution quality.