Science AI Journal
Engineering · Research topic

Open research questions in Infrastructure Maintenance and Monitoring

36 unresolved questions extracted from the limitations and future-work sections of 287 Infrastructure Maintenance and Monitoring papers in our library. Each links back to the study that raised it.

What the literature leaves open

  • Although various Explainable Artificial Intelligence (XAI) approaches have been proposed to highlight image regions influencing model predictions, qualitative visual inspection alone is insufficient for reliably evaluating the credibility of these explanations.

    Towards Global Interpretability: Evaluating XAI Metrics in Building Footprint Extraction · 2026 · DOI

  • This is attribu- ted to the effective feature extraction of the residual network and the key information focusing provided by the attention mechanism, enabling the model to capture core patterns even with limited data. Its superior performance under noise and limited data conditions, combined with strong generalization capability, positions it as a promising Wyniki eksperymentalne pokazują, że AM-ResNet osiąga do- skonałą dokładność prognozowania z R2 wynoszącym 0,790, MAE wynoszącym 3,21 MPa i MAPE wynoszącym 8,9%. Its advantage is particularly pronounced when training samples are scarce.

    Research on concrete strength prediction based on attention mechanism and residual network · 2026 · DOI

  • CiF establishes inspection of civil infrastructure, an elementary and seemingly easy perceptual task, as an open challenge that reveals fundamental weaknesses of present-day models trained predominantly on internet images, literally and figuratively highlighting cracks in the current foundation model paradigm.

    Cracks in the Foundation: A Civil Infrastructure Dataset to Challenge Vision Foundation Models · 2026

  • Future work will focus on extending validation to multi-source datasets covering diverse road marking types, pavement materials, and imaging conditions, as well as exploring adaptive indicator weighting strategies that can accommodate scene-specific variations without manual recalibration.

    <p>Automatic Assessment of Road Marking Surface Defects Using Computer Vision</p> · 2026 · DOI

  • Future research may focus on extending the framework to additional road condition indi- cators, improving domain generalization across different geographic environments, and optimizing the architecture for edge computing platforms to further enhance real-time road infrastructure monitoring and maintenance applications.

    A multi-scale transformer-enhanced YOLO framework for unified road damage detection and boundary-aware segmentation · 2026 · DOI

  • Therefore, future studies are recommended to integrate PKRMS analysis with structural pavement testing, traffic performance analysis, and drainage modeling to produce more comprehensive and accurate pavement maintenance strategies.

    Pavement Performance Evaluation Using the Provincial and Kabupaten Road Management System (PKRMS) Method on the Parengan–Lakardowo Road Section, Mojokerto Regency · 2026 · DOI

  • Future investigations will focus on optimizing these hybrid structures for edge-computing devices and expanding the dataset to include multi-scale damage types in diverse environmental settings to further enhance urban resilience. Future investigations may focus on incorporating more deep learning structures and optimization strategies to improve crack detection.

    A hybrid approach based on deep feature extraction and machine learning classification for structural damage detection in concrete structures · 2026 · DOI

  • An important limitation is that our analysis does not in- corporate the additional complexity of reporting standards such as EN13508-2 (CEN 2003), since SewerML labels are not based on this coding system.

    Multi-Class Sewer Defect Detection with Vision-Language Models · 2026 · DOI

  • The visualization results section appears incomplete ('it can be seen from Figure 5 that the original YOLOv11 model exhibited issues of false dete'), and no discussion of false positive/negative cases or failure modes is provided.

    Research on Pavement Crack Identification Method Based on Improved YOLOv11 · 2026 · DOI

  • A practical solution would be to design VLM outputs that combine textual explanations with real-time image highlighting of the perceived defect through bounding boxes or segmentation masks.

    Multi-Class Sewer Defect Detection with Vision-Language Models · 2026 · DOI

  • All models failed to detect Production Errors (PF) and performed poorly on Deformations (DE). Incorporating contextual information appears important to reduce hallucinations and improve accuracy for these defects.

    Multi-Class Sewer Defect Detection with Vision-Language Models · 2026 · DOI

  • Vegetation, branches, leaves, and shadows often produce false detections, making it difficult for AI models to distinguish real cracks from environmental interference.

    AI-Based Quantification of Crack Geometry on Retaining Walls from Mobile Earth-Observation Imagery · 2026 · DOI

  • 100. Best Applications 102. Immediate visuali- zation, high tem- poral resolution 106. Structured storage, analytical flexibility 110. Interoperability, modularity 114. Continuous moni- toring, predictive maintenance 103. Software-specific, limited scalability 104. Small to medium structures, pilot studies 107. Reduced real-time re- 108. Large infrastructure net- sponsiveness works 111. Additional computa- tional complexity 115. Bandwidth, security, visualization chal- lenges 112. Multi-platform projects, col- laborative workflows 116. Critical infrastructure, bridges, high-rise buildings This comparison illustrates that no single method is universally optimal; instead, in- tegration should be selected based on project scale, monitoring requirements, and stake- holder needs. Future research could explore hybrid approaches, combining the scalability of database-driven methods with the responsiveness of IoT streaming. 9.3 Strengths and Weaknesses of Current Research Current research in BIM–SHM integration demonstrates several notable strengths:  Enhanced Structural Safety: Continuous monitoring enables early detection of structural anomalies, reducing failure risk.  Predictive Maintenance and Lifecycle Optimization: SHM data integrated into BIM supports proactive interventions, lowering maintenance costs and material usage.  Sustainability Contributions: Integration facilitates energy optimization, material efficiency, and supports green certification initiatives.  Emerging AI Applications: Machine learning models improve prediction ac- curacy, anomaly detection, and automated BIM updates. However, several weaknesses remain, limiting broader adoption:  Standardization Deficiencies: Lack of universal protocols and semantic models inhibits interoperability across platforms and projects.  Data Management Challenges: High-frequency sensor data can overwhelm storage and analytics systems, creating bottlenecks.  Financial and Operational Constraints: High installation costs, mainte- nance, and training requirements limit adoption, especially for smaller pro- jects. 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 Fusion Journal of Engineering & Sciences 2025, 17, x https://doi.org/10.3390/xxxxx 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 Fusion Journal of Engineering & Sciences 2025 30 of 38  Human Factors: Limited expertise in both BIM and SHM systems creates skill gaps and affects decision-making.  Limited Real-World Validation: Many approaches are tested in pilot or con- trolled settings, with limited deployment in operational infrastructure.

    Integration of BIM and SHM For Sustainable Management of Existing Structures: A Systematic Review · 2026 · DOI

  • Traditional trial-and-error methods are time-consuming and costly, whereas existing data-driven models often lack robustness under limited data conditions and cannot provide actionable decision support for engineers.

    Performance Prediction and Optimization Design of Ultra-High Performance Concrete Based on Multi-Scale Residual Attention Convolution Network · 2026 · DOI

  • The results show that instance segmentation is a practical direction for field pavement imagery and aggregate crack-area estimation, while also exposing open challenges in annotation consistency, class imbalance, confounder rejection, and mask-level benchmarking.

    Pixel-Level Pavement Distress Assessment Using Instance Segmentation · 2026

  • 4124 Al-Mustafa & Ali One limitation is that SAM’s encoder operates at a fixed input size (1024×1024).

    PromptLessSAM: From Foundational Model to Domain Expert via Lightweight Decoder Adaptation for Crack Segmentation · 2026 · DOI

  • The paper lacks comparison with domain-specific pavement crack detection methods or discusses how the approach compares to manual inspection methods in terms of accuracy and cost-effectiveness.

    Research on Pavement Crack Identification Method Based on Improved YOLOv11 · 2026 · DOI

  • The paper does not address how the model performs on different types of pavement surfaces (asphalt vs. concrete) or varying environmental conditions, limiting its applicability across diverse real-world scenarios.

    Research on Pavement Crack Identification Method Based on Improved YOLOv11 · 2026 · DOI

  • The dataset is self-constructed but the paper provides limited details about dataset size, diversity, geographical variation, and seasonal variations in pavement conditions, which could affect model generalization.

    Research on Pavement Crack Identification Method Based on Improved YOLOv11 · 2026 · DOI

  • Two strategies could be pursued for dataset design: (i) targeted datasets that compensate for low-performing classes, or (ii) robustly sampled datasets that reflect field distributions.

    Multi-Class Sewer Defect Detection with Vision-Language Models · 2026 · DOI

  • The SewerML subset was not sampled to reflect the natural frequency distribution of classes in real-world inspections, so performance results should not be directly compared with larger, representative datasets.

    Multi-Class Sewer Defect Detection with Vision-Language Models · 2026 · DOI

  • Providing contextual cues in the prompts by linking georeferenced asset databases in an automated setting could improve the reliability of model outputs for Production Errors and Deformations.

    Multi-Class Sewer Defect Detection with Vision-Language Models · 2026 · DOI

  • A promising direction would be to fine-tune open-source models on recreated datasets that explicitly couple prediction, reasoning, and calibrated confidence to strengthen both interpretability and trustworthiness.

    Multi-Class Sewer Defect Detection with Vision-Language Models · 2026 · DOI

  • The study focuses on retaining walls in complex, vegetation-rich environments, suggesting potential limitations in generalization to other surface types or environmental conditions.

    AI-Based Quantification of Crack Geometry on Retaining Walls from Mobile Earth-Observation Imagery · 2026 · DOI

  • The approach provides crack metrics as a foundation for future retaining wall stability assessment and risk-informed infrastructure management, indicating that actual stability assessment models are not yet developed.

    AI-Based Quantification of Crack Geometry on Retaining Walls from Mobile Earth-Observation Imagery · 2026 · DOI

Most-cited papers in Infrastructure Maintenance and Monitoring

Most recent work

Find a gap in your own Infrastructure Maintenance and Monitoring sub-topic

This page shows what the Infrastructure Maintenance and Monitoring literature already flags as unresolved. To narrow it to your specific question, run the guided finder — it searches the gap library on demand and checks candidates against 250M+ OpenAlex works.

Open the Research Gap Finder →

Related topics in Engineering

Traffic and Road Safety
143 open questions
Spatial Cognition and Navigation
95 open questions
Smart Cities and Technologies
89 open questions
BIM and Construction Integration
87 open questions
Design Education and Practice
86 open questions
Engineering Education and Curriculum Development
82 open questions
Digital Transformation in Industry
81 open questions
Perovskite Materials and Applications
78 open questions
36 open questions have been extracted from the limitations and future-work passages of 287 Infrastructure Maintenance and Monitoring papers in our library. Each one below links back to the study that raised it, so you can read the original claim in context.

Tools for your next paper

Gaps & Literature SearchFind open questions in any field — 120,000+ open gaps mined from our own library.Pre-CheckIs your paper ready? A Tier 1–5 acceptance probability in seconds.Journal FinderA ranked shortlist of target journals from a 17,500-venue index.Duplicate Publication CheckerCheck prior-publication & salami-slicing risk across 8 sources.Citation GeneratorFreeA DOI or title → APA, MLA, Vancouver, Chicago, BibTeX, RIS. No signup.Graphical Abstract MakerFreeTurn your findings into a colorblind-safe graphical abstract.

Compare the category — Honest roundups of the AI research tools, ours listed alongside the alternatives.

Command palette

Jump anywhere, run any action.