Real-Time Pothole Detection and Monitoring Using Machine Learning and IoT Integration for Enhanced Road Safety

Potholes are a serious traffic danger that can cause accidents and present difficulties for vehicles, especially during the rainy season. Conventional techniques for detecting potholes, like manual inspection and vibration-based sensors, are expensive, time-consuming, and fraught with danger. This study suggests a non-invasive, real-time pothole detection system that makes use of the YOLOv8 algorithm, a cutting-edge convolutional neural network (CNN) model for object detection, to overcome these constraints. YOLOv8 predicts bounding boxes and item classes in road photos simultaneously, improving accuracy and responsiveness. With the help of this system, which can be integrated into self-driving and autonomous vehicles, potholes may be identified and visually highlighted in real-time, making navigation safer. A variety of datasets with 720 x 720-pixel resolution photos that captured different pothole scenarios in real-world road conditions were used to train, test, and validate the model. The outcomes show how well YOLOv8 detects and categorizes potholes, offering a reliable option for pothole care and accident avoidance. The effectiveness of CNN-based methods in improving road surface monitoring and safety is highlighted by this study.