AGMAMBA: A LIGHTWEIGHT ADAPTIVE GUIDED STATE SPACE MODEL FOR PIXEL-LEVEL CRACK SEGMENTATION IN CIVIL INFRASTRUCTURE

Accurate pixel-level segmentation of crack regions is essential for the inspection and maintenance of civil infrastructure such as bridges, pavements, and buildings. Early detection of structural cracks enables timely maintenance actions and reduces the risk of structural failure. However, existing segmentation approaches often struggle to simultaneously capture fine crack textures and suppress complex background noise, which can significantly degrade detection accuracy. Additionally, many high-performing models require large computational resources, limiting their applicability in real-time infrastructure monitoring systems.

Challenges in Existing Crack Segmentation Methods

Current deep learning-based crack detection methods frequently encounter two primary challenges. First, cracks typically exhibit thin, irregular, and discontinuous patterns, making it difficult for models to dynamically capture their texture and morphology. Second, background elements such as shadows, stains, and surface roughness can easily be misidentified as cracks. These challenges often lead to reduced segmentation accuracy or excessive computational complexity when attempting to improve feature extraction.

AGMamba Architecture and Lightweight Design

To overcome these limitations, this study introduces AGMamba, a lightweight adaptive guided state space model designed for efficient crack segmentation. The architecture focuses on capturing crack texture features while minimizing redundant background information. With only 3.40 million parameters and 20.99 GFLOPs, AGMamba achieves a strong balance between computational efficiency and segmentation performance, making it suitable for practical civil infrastructure inspection applications.

Crack Perception Module (CPM)

A key component of the proposed model is the Crack Perception Module (CPM), which integrates two complementary mechanisms:

• Adaptive Guided Scanning Strategy (AGSS2D) – prioritizes crack regions during feature scanning to improve the efficiency of texture extraction.

• Selective Key Clue Modeling (SKCM) – selectively aggregates information from critical crack edges and structural features.

Together, these modules allow the network to focus on meaningful crack patterns while reducing the influence of irrelevant background features.

Frequency-Domain Feature Perception

To further enhance segmentation performance, the model incorporates a High-Low Frequency Feature Perception (HLFP) strategy and a Frequency-Domain Segmentation Head (FDSH). These components analyze differences between high-frequency crack textures and low-frequency background patterns. By leveraging frequency-domain information, the framework effectively suppresses background interference and improves crack boundary detection accuracy.

Experimental Results and Performance Evaluation

Extensive experiments were conducted on four public crack datasets, demonstrating that AGMamba consistently outperforms existing state-of-the-art (SOTA) segmentation models. On the Crack500 dataset, the proposed model achieved an F1 score of 0.7622 and an mIoU of 0.7808, representing improvements of 1.41% and 1.21%, respectively, over previous SOTA methods. These results confirm the model’s ability to achieve high segmentation accuracy while maintaining low computational cost, making it highly suitable for automated infrastructure inspection systems.

Global Civil Engineering Awards

🔗 Nominate now! 👉 https://civilengineeringawards.com/award-nomination/?ecategory=Awards&rcategory=Awardee

🌐 Visit: civilengineeringawards.com

📩 Contact: contact@civilengineeringawards.com

#BridgeInspection
#PavementMonitoring
#AIinCivilEngineering
#StructuralSafety
#AutomatedInspection
#DigitalInfrastructure
#LightweightAI
#EngineeringVision
#CivilEngineeringResearch
#SmartMaintenance

 

BRIDGE DIGITAL TWINS AND THE ROLE OF LOAD TESTING IN LIFECYCLE MANAGEMENT

Bridge digital twins represent a transformative approach in bridge engineering, enabling the creation of virtual replicas that mirror the physical structure throughout its lifecycle. Originating from advancements in other industries, digital twin technology integrates real-world data with computational models to support monitoring, analysis, and decision-making. In bridge applications, digital twins promise enhanced safety, predictive maintenance, and optimized asset management, making them a critical component of next-generation infrastructure systems.

Concept of Digital Twins in Bridge Engineering

A bridge digital twin combines geometric information, material properties, sensor data, and operational conditions into a unified virtual model. By linking physical bridges with Building Information Modeling (BIM) and Finite Element (FE) models, engineers can simulate structural behavior under varying loads and environmental influences. This integration enables continuous assessment of performance, allowing infrastructure owners to move from reactive maintenance toward proactive management strategies.

Load Testing as the Birth of the Digital Twin

The study emphasizes that bridge load testing marks the “birth” of the digital twin. During this phase, controlled loads are applied to the actual bridge to measure structural responses such as deflection, strain, and vibration. This process provides high-quality empirical data that cannot be obtained at later stages with the same reliability. Consequently, load testing offers a unique opportunity to calibrate digital models so that they accurately represent real structural behavior.

Updating BIM and Finite Element Models

Accurate digital twins depend on well-calibrated BIM and FE models. Load test data enables engineers to validate assumptions regarding stiffness, boundary conditions, and material properties. By updating these models with measured responses, discrepancies between theoretical predictions and actual performance can be minimized. This calibration ensures that the digital twin remains a trustworthy tool for structural analysis, safety evaluation, and performance forecasting.

Application During the Operational Phase

Once established, the digital twin supports the bridge throughout its service life. During operation, it can be continuously updated with monitoring data to detect anomalies, assess damage, and evaluate the effects of aging or environmental changes. This capability allows engineers to predict future performance, schedule maintenance efficiently, and extend service life while maintaining safety standards. Thus, the digital twin becomes an active management system rather than a static model.

Case Study of a Post-Tensioned Concrete Bridge

The concept is demonstrated through modeling and load testing of a real post-tensioned concrete bridge. Post-tensioning introduces complex stress distributions and structural behavior, making accurate modeling particularly important. The case study illustrates how field measurements obtained during testing can refine computational models and establish a reliable digital twin. This example confirms the feasibility and practical value of integrating testing, modeling, and digital technologies in modern bridge engineering.

🏗️ Civil Engineering Awards  

👉 Visit our Website: civilengineeringawards.com

#InfrastructureInnovation

#StructuralAnalysis

#EngineeringSimulation

#SmartBridges

#AssetManagement

#StructuralSafety

#InfrastructureLifecycle

#FutureOfConstruction

 

SLAM-CENTRIC FRAMEWORK FOR PRECISE AND PLATFORM-AGNOSTIC ROBOT-AIDED INFRASTRUCTURE INSPECTION

Robot-aided inspection has emerged as a promising solution for enhancing safety, efficiency, and objectivity in infrastructure assessment. However, existing approaches often suffer from inconsistent mapping accuracy, unreliable defect measurements, and platform-specific system designs that limit scalability. This study investigates whether a SLAM-centric (Simultaneous Localization and Mapping) framework can overcome these limitations and enable precise, repeatable, and platform-agnostic visual inspections across diverse infrastructure environments.

Integrated Lidar–Camera–Inertial SLAM Architecture

The proposed framework integrates lidar, camera, and inertial measurement unit (IMU) data within a unified SLAM pipeline to ensure robust localization and mapping under real-world conditions. Multi-sensor fusion enhances pose estimation accuracy and resilience to environmental challenges such as lighting variation, occlusions, and geometric complexity. By centering the inspection workflow around high-fidelity SLAM, the system establishes a reliable spatial reference for defect mapping and measurement, independent of the robotic platform employed.

Offline Trajectory Refinement and Inspection Map Generation

To further improve mapping precision, the framework incorporates offline trajectory refinement, reducing drift and cumulative localization errors commonly observed in real-time SLAM systems. Refined trajectories enable the generation of dense and geometrically consistent inspection maps. These maps serve as a unified spatial representation where inspection data can be consistently overlaid, facilitating repeatable assessments and longitudinal monitoring of infrastructure assets.

Automated Defect Extraction and 3D Ray-Tracing Projection

Visual defect detection is performed through image-based analysis, extracting cracks, spalls, and surface anomalies from captured imagery. A 3D ray-tracing technique projects detected defects into the unified inspection map, ensuring accurate spatial localization and dimensional quantification. This method allows precise measurement of defect size, orientation, and position within the 3D structure, significantly improving reliability compared to traditional qualitative or manual inspection methods.

Validation in Real-World Scenarios

Experimental validation in real-world environments demonstrates that the SLAM-centric framework produces accurate defect localization, consistent dimensional measurements, and high-density inspection maps. The platform-agnostic design ensures adaptability across different robotic systems, including ground vehicles, aerial drones, and climbing robots. The repeatability and robustness of the approach confirm its suitability for practical infrastructure inspection applications.

Implications for Long-Term Monitoring and Automation

By providing an end-to-end solution for robot-aided inspection, the proposed framework enables faster, safer, and more objective infrastructure assessments. The release of datasets and open software tools establishes a foundation for future research in long-term defect monitoring, inspection automation, and predictive maintenance. This SLAM-centric paradigm represents a significant step toward intelligent, data-driven infrastructure management systems.

🏗️ Civil Engineering Awards  

👉 Visit our Website: civilengineeringawards.com

#InspectionAutomation

#SmartInfrastructure

#CivilEngineeringTechnology

#DroneInspection

#RoboticsInConstruction

#InfrastructureSafety

#AIforEngineering

#PredictiveMaintenance

#DigitalTwin

 

 

NOVEL SELF-POWERED SENSOR (NSPS) FOR INTELLIGENT STRUCTURAL HEALTH MONITORING OF CIVIL INFRASTRUCTURE

 

Structural Health Monitoring (SHM) plays a critical role in ensuring the safety, durability, and serviceability of civil infrastructure such as bridges, buildings, and transportation systems. Conventional monitoring systems often depend on external power supplies and wired data transmission, limiting their scalability and long-term reliability. This study introduces a Novel Self-Powered Sensor (NSPS) specifically designed for civil structures, integrating self-energy harvesting, low-power wireless communication, and intelligent sensing capabilities. The proposed system addresses the limitations of traditional SHM by enabling sustainable, long-term, and autonomous infrastructure monitoring.

System Architecture and Core Technologies

The NSPS integrates three major technological components: environmental energy harvesting, ultra-low-power wireless data transmission, and intelligent sensing modules. The energy harvesting unit captures ambient environmental energy—such as vibration, solar, or thermal energy—and converts it into usable electrical power. The low-power wireless transmission system enables large-scale deployment across infrastructure networks without extensive cabling. Intelligent sensing algorithms process structural performance data efficiently, ensuring accurate detection of stress, deformation, and environmental variations under complex operational conditions.

Energy Harvesting Optimization and Power Management

A central innovation of the NSPS lies in its self-powered functionality. By fine-tuning the sensor design, optimal energy conversion efficiency is achieved from the harvesting unit, ensuring continuous operation even under variable environmental conditions. Advanced power management strategies regulate energy storage, consumption, and transmission cycles to maintain stable performance. This optimization enables the sensor to operate over extended periods without battery replacement, significantly reducing maintenance costs and enhancing the sustainability of monitoring systems.

Wireless Communication and Large-Scale Deployment

The NSPS employs large-scale, low-power wireless data transmission protocols that facilitate real-time structural performance monitoring across extensive infrastructure networks. This approach reduces installation complexity and allows flexible sensor placement in remote or hard-to-access areas. Compared to conventional wired systems, the wireless architecture improves coverage, scalability, and data accessibility, supporting integrated monitoring platforms for smart infrastructure management.

Bridge Case Study and Monitoring Strategy Development

To validate the practicality and effectiveness of the NSPS, a case study is conducted on an operational bridge structure. A monitoring scheme is developed based on the structural principles and load-bearing characteristics of the bridge. The sensor deployment strategy considers key stress zones, dynamic load responses, and environmental exposure conditions. Field testing demonstrates the system’s reliability in real-world scenarios, confirming its ability to continuously collect and transmit high-quality data while maintaining energy autonomy.

Sustainability, Performance Evaluation, and Future Applications

When compared to traditional structural monitoring techniques, the NSPS demonstrates significant improvements in sustainability, operational efficiency, and long-term reliability. The elimination of frequent battery replacement and reduced wiring requirements contribute to lower lifecycle costs and environmental impact. This innovative self-powered monitoring solution lays a strong foundation for future advancements in intelligent transportation systems and smart infrastructure. Further research may focus on multi-energy harvesting integration, AI-based damage prediction, and large-scale implementation across diverse civil engineering applications.

🏗️ Civil Engineering Awards  

👉 Visit our Website: civilengineeringawards.com

#InfrastructureSafety
#SmartBridges
#StructuralMonitoring
#IoTSensors
#EngineeringResearch
#DigitalInfrastructure
#ResilientStructures
#TransportationEngineering
#SHMTechnology
#GreenEngineering

Intelligent Infrastructure Crack Detection Using MSEDBO-Optimized Deep Learning

 

Infrastructure surface crack detection is a vital task in structural health monitoring, directly influencing the safety, durability, and serviceability of civil engineering assets. Although deep learning methods have achieved notable success in automated crack detection, their performance is often constrained by inefficient hyperparameter tuning, susceptibility to local optima, and suboptimal feature extraction. This study addresses these limitations by proposing an intelligent optimization-driven crack detection framework.

Limitations of Conventional Deep Learning-Based Crack Detection

Traditional deep learning models rely heavily on manual or heuristic-based hyperparameter selection, which can lead to unstable training outcomes and reduced generalization performance. Moreover, commonly used optimization techniques may become trapped in local optima, resulting in inaccurate crack localization and increased false positive rates, particularly when dealing with complex backgrounds and diverse infrastructure materials.

Multi-Strategy Enhanced Dung Beetle Optimizer (MSEDBO)

The proposed framework integrates a Multi-Strategy Enhanced Dung Beetle Optimizer (MSEDBO) to systematically optimize critical parameters within the crack detection pipeline. MSEDBO incorporates Latin Hypercube Sampling with elite population initialization, an improved sigmoid-based nonlinear control factor, sine–cosine algorithm integration, and multi-population mutation strategies. These enhancements collectively strengthen global exploration and local exploitation capabilities.

Integration with Deep Learning Models

By embedding MSEDBO into deep learning-based crack detection models, the framework enables adaptive optimization of network parameters and feature extraction processes. This synergy improves convergence behavior, enhances robustness against local optima, and ensures efficient learning across varying crack patterns and surface conditions in civil infrastructure.

Experimental Validation and Benchmark Datasets

The proposed approach was validated using multiple benchmark datasets, including CrackTree200, CFD, GAPs, and SDNET2018, covering a wide range of materials such as concrete pavements, asphalt roads, and bridge surfaces. Comparative experiments demonstrate that the MSEDBO-optimized framework consistently outperforms conventional optimization algorithms and baseline deep learning models.

Performance Gains and Practical Implications

Results show significant improvements, including an 8.7% increase in detection accuracy, a 12.3% improvement in precision, and a 15.6% reduction in false positive rates. The framework maintains computational efficiency while effectively avoiding local optima, making it well suited for real-world deployment. This research advances intelligent infrastructure monitoring by providing a robust optimization strategy to enhance the reliability and accuracy of automated crack detection systems.

🏗️ Civil Engineering Awards  

👉 Visit our Website: civilengineeringawards.com

#AIinCivilEngineering

#AutomatedInspection

#ConcreteCracks

#RoadSurfaceMonitoring

#BridgeInspection

#MachineLearning

#EngineeringOptimization

#DigitalInfrastructure

#SustainableInfrastructure

#CivilEngineeringResearch

Transformer-Based Intelligent Defect Detection in Civil Infrastructure

Detecting structural and functional defects in large-scale civil infrastructure during operational stages is critical for ensuring safety, serviceability, and sustainability. Traditional inspection methods are often labor-intensive, subjective, and inefficient for large-scale systems. With the rapid evolution of artificial intelligence, deep learning has emerged as a powerful tool for intelligent defect detection. Recently, Transformer-based self-attention models have gained prominence as effective alternatives to convolutional neural networks (CNNs), offering superior capability in modeling long-range dependencies and parallel computation. This research-oriented survey systematically explores the integration of Transformer architectures into civil engineering defect detection applications.

Evolution from CNNs to Transformer Models

Conventional CNN-based approaches have demonstrated strong performance in localized feature extraction for defect detection tasks; however, their limited receptive fields restrict the modeling of global structural relationships. Transformers overcome these limitations through self-attention mechanisms that capture global contextual information across large-scale datasets. This paradigm shift has motivated civil engineering researchers to adopt Transformer models for infrastructure monitoring, enabling more comprehensive understanding of spatial and temporal defect patterns.

Transformer Architectures for Engineering Defect Detection

This survey reviews more than 40 Transformer-based engineering defect detection algorithms, highlighting key architectural variants such as Vision Transformers (ViT), hybrid CNN-Transformer models, and hierarchical attention frameworks. These architectures are tailored to handle high-resolution images, sensor data, and multimodal inputs commonly encountered in civil infrastructure monitoring. The adaptability of Transformer models allows effective feature learning across complex structural geometries and diverse defect manifestations.

Application Scenarios in Civil Infrastructure

Transformer-based defect detection methods have been extensively applied to critical civil infrastructure systems, particularly roadways, tunnels, and bridges. In roadway monitoring, Transformers enable accurate detection of cracks, potholes, and surface degradation. Tunnel inspection applications benefit from long-range dependency modeling in low-light and complex environments, while bridge monitoring leverages attention mechanisms to identify structural anomalies across large spans and interconnected components.

Challenges and Limitations

Despite their advantages, Transformer-based approaches face several challenges in civil engineering applications. These include high computational costs, large data requirements, limited labeled datasets, and difficulties in real-time deployment. Additionally, domain adaptation across different infrastructure types and environmental conditions remains a significant research challenge. Addressing these issues is essential for practical, scalable adoption in real-world infrastructure systems.

Future Research Directions

Future development of Transformer-based intelligent detection systems should focus on lightweight architectures, self-supervised learning, multimodal data fusion, and domain-specific model optimization. Integrating Transformers with digital twins, edge computing, and real-time monitoring systems holds promise for next-generation smart infrastructure management. This survey provides foundational insights and reference strategies to guide researchers in advancing intelligent defect detection within civil engineering.

🏗️ Civil Engineering Awards  

👉 Visit our Website: civilengineeringawards.com

#RoadwayEngineering

#VisionTransformer

#AIinEngineering

#SelfAttention

#IntelligentSystems

#DigitalInfrastructure

#EngineeringResearch

#StructuralSafety

#AutomationInCivil

#FutureEngineering

#SustainableInfrastructure