The increasing prevalence of autonomous vehicles on our roads necessitates the development of robust systems to ensure their safe navigation, particularly in the face of deteriorating road conditions. Potholes pose significant hazards, leading to vehicular damage and accidents, underscoring the need for timely detection and remediation. This paper introduces a novel approach to pothole detection by implementing the YOLOv4 (You Only Look Once version 4) deep learning algorithm, which excels in real-time object detection tasks. Our method aims to enhance the accuracy and efficiency of pothole identification, enabling autonomous vehicles to better navigate urban environments while promoting overall road safety and reliability.