Publications

Abstract : Potholes are perennial problem across the world. They are dangerous and causes road accidents. The potholes often begin as imperceptible microscopic cracks in the road surface. The main reason behind these potholes formation are bad weather, poor drainage maintenance and construction, high traffic, which causes that surface to loosen and damage. This issue has to be reported to the roadways department by identifying cracks and potholes on the roads. This would help them to resolve the issue effectively and prevent the potholes from causing fatal damages. Image Processing helps in detecting the Potholes, and the predictions made are quite accurate. By utilizing the images of the cracked potholes, asphalt failures and fatigue potholes on the roads, a system is proposed to identify the potholes. The images are fed into the YOLO (You Only Look Once) algorithm, which is used to detect the cracks leading to deep potholes. YOLOv2 helps us predict potholes in real time object detection at the rate of 45 FPS (Frames Per Second).