This dataset contains 2.7k images and is mainly used to detect potholes, cracks and open manholes on the road. The dataset has been enhanced to increase the diversity and robustness of the data. The images are divided into training and validation sets and are divided into 3 different categories:

• Potholes: Level 0
• Crack: Level 1
• Open manholes: Category 2

The dataset contains bounding box annotations in .txt files formatted for YOLOv8s, ensuring compatibility for model training. It has separate folders for each category and contains training, valid, and all category folders for easy access and custom enhancement. The dataset is intended for further model training, testing, and custom enhancement tasks related to road safety and infrastructure detection.