| Summary: | In rapidly urbanizing regions, maintaining road infrastructure integrity is a critical challenge due to increasing vehicular stress and environmental factors. This study introduces the Road Crack Dataset-International Islamic University Malaysia (RCD-IIUM), designed to enhance road pavement infrastructure management in Malaysia. Employing advanced data collection technologies, including high-resolution digital imaging, the dataset captures detailed anomalies in road surfaces, laying the groundwork for robust infrastructure analysis. The utility and efficacy of the RCD-IIUM dataset were evaluated through the deployment of three deep learning models: Customized YOLOv7, YOLOv8X-SEG, and an Advanced Hybrid Deep Learning Model. These models were tested for their ability to detect and classify road cracks using metrics such as precision, recall, F1-score, and overall accuracy. Results indicated that the YOLOv8X-SEG model outperformed others, demonstrating higher accuracy of 90% and F1-score of 95%. The Customized YOLOv7 model achieved a precision of 93 %, recall of 91.58%, and overall accuracy of 88%. The Advanced Hybrid Deep Learning Model achieved a precision of 88%, recall of 89%, F1-score of 88.5%, and overall accuracy of 85%, further validating the robustness of the dataset. The dataset not only bolsters road pavement maintenance strategies but also supports data-driven decision-making for urban planning and policymaking. © 2024 IEEE.
|
|---|