Development of a Pothole Detection System Using Deep Learning Techniques and Depth Estimation

Our paper presents a simple and effective method for detecting and estimating the depth of potholes, a common problem in urban areas that poses a serious threat to drivers and causes extensive vehicle damage, leading to increased maintenance costs. We have implemented YOLOv8 which has not been implemented to detect potholes. We have proposed a unique and effective method to calculate the depth of potholes. Despite the prevalence of potholes, research on estimating their depth has been limited. To address this issue, we propose a pothole detection system using the You Only Look Once version 8 (YOLOv8) algorithm and You Only Look Once version 5 (YOLOv5) algorithm, both are state-of-the-art object detection algorithms. Our system is evaluated on a dataset of road images, and we demonstrate its superior performance compared to existing pothole detection techniques. Our approach offers a time-efficient and cost-effective solution to detecting and estimating the depth of potholes using deep learning techniques.