Aug 2026· Bulletin of Computational Intelligence· Vol 2, pp. 235-251· 0 citations· 63 references
TL;DR
The findings from the review indicate that YOLO and CNN were extensively employed by researchers, however in recent times, vision transformers gained popularity among researchers and pavement engineers, leading to safer, more resilient, and sustainable transportation infrastructure.
Abstract
Pavements are essential elements of transportation networks and are instrumental to the growth and development of a country. To harness the full potential of this infrastructure, it is necessary to maintain it in proper structural and functional conditions. Conventional techniques, such as manual inspection can be used to determine the functional conditions of the pavement. However, these techniques have different shortcomings, including high monitoring costs, time consumption and requirement of traffic control during pavement surveying. These shortcomings can be mitigated by utilizing various Artificial Intelligence (AI) techniques such as machine learning and deep learning, for pavement surface inspection. This study aims to systematically review state-of-the-art deep learning techniques such as vision transformer model for pavement distress detection. This research contributes to the knowledge body by summarizing the application of vision transformer model for pavement distress detection. Moreover, this research provides a brief description of the application of other deep learning architectures including You Only Look Once (YOLO), and Convolutional Neural Networks (CNNs) for pavement distress detection. Deep learning techniques can autonomously detect various types of pavement distress including longitudinal and transverse cracks, rutting, faulting, patching, shoving, raveling and potholes from the pavement surface. The research was performed in three different steps namely: keyword selection, identification of databases, literature acquisition and summarizing the literature based on different deep learning models. The findings from the review indicate that YOLO and CNN were extensively employed by researchers, however in recent times, vision transformers gained popularity among researchers and pavement engineers. Overall, this study highlights the critical role played by different deep learning techniques in transforming pavement monitoring, leading to safer, more resilient, and sustainable transportation infrastructure.
This review has organized improvements in SHM along the lines of vibration-based anomaly detection, vision-based defect recognition, and multi-modal data fusion along the lines of vibration-based anomaly detection, vision-based defect recognition, and multi-modal data fusion.
Bellal Mia, Md Umar Faruk, M. Hasan et al.· Scientia. Technology, Scienc...· 0 citations
A Yolov8-MED design integrate MHSA, ECA and DCNv3 to enhance the model’s ability to capture fine crack details and improve detection accuracy, to enhance the crack location accuracy and keep the computational cost simultaneously.
Mohammed Al-Soswa, Mohammed Al-Mahbashi, Zhaoyun Sun et al.· Computing· 0 citations
Effective monitoring of pavement conditions plays an important role in ensuring road safety and optimizing maintenance. Traditional inspection methods are labor-intensive and limited in coverage, motivating the use of remote sensing and deep learning for scalable alternatives. This study proposes a novel framework for classifying pavement type and condition by integrating medium-resolution optical imagery from Planet satellite and synthetic aperture radar (SAR) data from Sentinel-1. A deep learning dataset was constructed by combining eight bands and labeling them using IRI data collected over multiple years. Two semantic segmentation models, U-Net and DeepLabV3, were trained and evaluated. Also, a series of band-drop scenarios was tested to assess the importance of each spectral and radar band. Results showed that DeepLabV3 outperformed U-Net in both overall accuracy and weighted
F
1
-score. Particularly, DeepLabV3 was more successful in identifying fair and poor asphalt pavements, which are critical for proactive maintenance decisions. The analysis illustrates that combining SAR with optical data consistently improved results compared to using either source independently. These findings demonstrate the feasibility and effectiveness of using multisource satellite imagery with deep learning for scalable, cost-efficient pavement condition assessment. The proposed framework offers transportation agencies a practical solution for enhancing infrastructure monitoring and data-driven decision-making.
Mahyar Shahri, S. Sonny Kim· Journal of Transportation En...· 0 citations
Pavement raveling is characterized by the progressive loss of surface aggregates and poses significant challenges to road durability and safety. The early detection and monitoring of this disease are crucial for implementing timely maintenance interventions. This study adopts a transfer learning approach for automated raveling detection in asphalt pavements. A comprehensive
in situ
raveling dataset comprising 530 images and 215 normal pavement images was collected using the LS-40 portable three-dimensional surface analyzer. Following histogram equalization for noise reduction, the dataset was augmented to 1600 images through mirroring and rotation techniques. The transfer learning was fine-tuned on four base convolutional neural network structures: the VGG16, the EfficientNet-B0, the InceptionV3, and the RegNet. The models were trained using standardized parameters, including an input size of 224 × 224, a batch size of 32, and 200 epochs, with regularization and dropout techniques applied to mitigate overfitting. A comparative analysis of optimization algorithms showed that RMSprop outperformed both SGD and Adam for this specific task. Transfer learning significantly enhanced the performance of all models, with the EfficientNet-B0 achieving outstanding results—achieving both high accuracy (99.7%) and low energy consumption in raveling detection. The selected models achieved AUC values exceeding 0.95, while the transfer learning significantly reduced training time and enhanced feature extraction capabilities, as confirmed through convolutional layer visualization. These findings establish the EfficientNet-B0 as the optimal structure for practical deployment in automated pavement inspection systems, providing a robust foundation for intelligent infrastructure maintenance strategies.
Juanjuan Wen, Yi Jiang, Yi Peng· Frontiers in Materials· 0 citations
This review provides a state-of-the-art survey of image-based automation techniques for sewer inspection, focusing on the progression from traditional rule-based image processing to advanced deep learning models. The core computer vision tasks are examined, including defect classification, object detection, and semantic segmentation, with an emphasis on their capabilities and limitations. Recent trends, such as the use of multilabel classification to detect multiple defects in a single image and transformer-based models for better understanding of visual context, are also discussed. The review also explores the emerging role of three-dimensional (3D) point cloud analysis and 3D reconstruction techniques in enhancing spatial understanding and defect localization. Despite notable advances, several challenges remain, such as the scarcity of standardized datasets, limited real-world validation, and the absence of fully automated deployable systems. To address these gaps, the paper recommends promoting open-access benchmark datasets; encouraging the open sharing of code and models; the adoption of data-efficient learning approaches such as unsupervised, semisupervised, self-supervised, and transfer learning to reduce reliance on large annotated datasets; the use of lightweight and deployable model architectures; and the validation of AI-driven models in real-world sewer inspection environments. These directions are essential to transforming current research innovations into practical, intelligent inspection systems.
Pavement performance prediction aims to forecast the future condition of pavement as accurately as possible, enabling proactive planning and optimized maintenance and rehabilitation (M&R) interventions. Pavement failure has traditionally been predetermined during the design stage when performance indicators such as the international roughness index (IRI), cracking, and other distresses exceed predefined thresholds. While such projections may be useful during material-selection phases, pavements often experience varying deterioration patterns over their service life, making original forecasts less reliable. To overcome these challenges, we propose to use advanced predictive models based on machine-learning techniques while leveraging text embeddings extracted from pretrained models. This study specifically presents two IRI prediction models: a long short-term memory model for forecasting IRI during long-term deterioration, and an artificial neural network model for predicting post-maintenance IRI. Our study further experiments with the semantic encoding power of different large language models, including transformer models, to provide numerical representations of raw maintenance logs. We show the importance of precise parameter selection for text-embedding models along with readily available pavement-design input parameters passed to a customized artificial neural network (ANN) model for post-maintenance IRI prediction. The proposed framework supports integration of a wide variety of maintenance actions, including complex scenarios involving multiple maintenance applications. We trained the models using data extracted from the long-term pavement performance (LTPP) database and obtained
R
2
scores of 93% and 86% for IRI long-term deterioration and post-maintenance test datasets, respectively. These results underscore successful application of intelligent models across the pavement’s entire service life.
K. S. Oguntoye, H. Ceylan, Berk Gulmezoglu et al.· Transportation Research Reco...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
