Pothole Detection and Prediction Using Deep Learning with CNN and YOLOv8

Potholes cause great danger in driving and to the lives of passengers. Therefore, it is essential for the accurate prediction and detection of potholes and it also saves the lives of many people. Pothole objects are difficult to accurately detect due to their regular shape in challenging, and dynamic road environments. This research study aims to improve the accuracy of detecting and predicting potholes using deep learning methods (such as ResNet and YOLO). The deep learning method is verified as effective and efficient in object detection. The results ratify that the proposed models have the feasibility to detect pothole occurrence.