Smart Pothole Detection and Traffic Sign Identification for Indian Roads: A Machine Learning Approach Using Yolov11

Road travel is one of the most common modes of transportation internationally, with over 60 % of the population relying on personal or shared vehicles. Regular maintenance of roads ensures their safety and functionality. Survey results indicate that potholes significantly contribute to various incidents on the roads. As a consequence of the aforementioned issues, roads deteriorate prematurely or poor practices in construction are leading to the formation of potholes that leads to accidents. According to the Ministry of Road Transport and Highways India (MoRTH) data, the total number of road accident deaths due to potholes in 2019 and 2020 stood at 2,140, and 1,471, respectively. This research proposes a methodology for pothole detection using the YOLOv11 object detection algorithm applied to road images achieving 80.4 % of precision on our dataset with other 19 classes in a multi class setting. Our dataset includes road images featuring potholes and 19 additional classes such as speed Limit, automobile, no parking, accident prone area and traffic signals. The proposed approach holds the potential for real-time pothole detection offering prospects to enhance road safety and diminish maintenance costs on roads. While numerous pothole detection models rely on accelerometers without utilising images or videos in conjunction with machine learning techniques, there is a need for more models that exclusively employ machine learning for pothole detection. The outcomes of our research demonstrate the superior performance of our proposed model compared to existing pothole detection techniques.