Detection of pothole for repair works of asphalt flexible pavement optimization using YOLO

In Malaysia, the government has consistently allocated a significant amount of budget in improving and maintaining country's road service quality. To ensure best-value maintenance works, it is essential to carry out regular pavement distress surveys for efficient management planning. Pothole long-standing issue most highlighted with many lives have been claimed over years. A surveyor detects by manually inspecting surface, therefore tedious, time-consuming, laborious hazardous process. This study proposes new method assist surveyors detecting potholes using deep learning approach. Road surface images are captured from federal roads Kedah. The LabelImg app used as tool label defining bounding box around each pothole. Finally, labelled presented called You Only Look Once (YOLO), train network given images. YOLO was found achieve promising results under various conditions, an accuracy up 80%. area YOLO's also beneficial estimating pothole volume repair work.