ML-Driven Detection and Quantification of Potholes for Safer Roads

Author(s)	Prof. Dr. Miruna Joe Amali S, Ms. Vaishali K S, Ms. Yogeshwari N O
Country	India
Abstract	Road safety remains a critical challenge in urban environments where potholes contribute significantly to vehicle damage and accidents. Traditional manual inspection methods are time-consuming, expensive, and often lack real-time precision. This paper proposes a Machine Learning-driven approach for real-time detection and quantification of potholes using the YOLOv5 object detection model. The system introduces severity classification (Minor, Moderate, Severe) and monocular depth estimation to assess the pothole’s impact level. Additionally, speed and distance-aware alerting mechanisms ensure that drivers receive warnings only when risk is immediate, reducing false alerts in static traffic. The proposed system demonstrates strong potential for integration into smart city infrastructures, offering valuable insights for both drivers and road maintenance authorities.
Keywords	YOLOv5, Pothole Detection, Severity Classification, Depth Estimation, Computer Vision, Smart City, Real-Time Alert System
Field	Computer > Artificial Intelligence / Simulation / Virtual Reality
Published In	Volume 7, Issue 5, September-October 2025
Published On	2025-10-25
DOI	https://doi.org/10.36948/ijfmr.2025.v07i05.58751

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research