Potholes Detection System Utilizing Visual Geometry Group-16 (VGG16)

Potholes pose a great threat to vehicle integrity and road safety. Roads play a crucial role in our economy and serve as a major platform for transportation, which is why adequate maintenance of the roads is crucial to keep them functional. And due to use of poor construction material, potholes appear, and the road gets damaged, which makes them unsafe and causes many unwanted incidents. Deep learning has emerged as a promising approach for pothole detection. This study is deployed to test and detect potholes using deep learning techniques, specifically VGG16 and dives into deep learning for pothole detection, training, and performance analysis on road images. It shows deep learning's effectiveness for pothole detection and improved road safety. Therefore, this paper demonstrates significant results in detecting potholes that can benefit the government to detect potholes in real time and ensuring safety.