• Version
• Download 5
• File Size 382.14 KB
• File Count 1
• Create Date 21 May 2025
• Last Updated 21 May 2025

“Automatic Pothole Detection on Road Surface”

Authors:

Sneha Sonawane, Shraddha Rasal, Shrisha Patil, Kashish kute, Prof. V.J. Bodake

Computer Department Loknete Gopinathji Munde Institute of Engineering Education & Research

ABSTRACT - Road safety and infrastructure maintenance are growing concerns in developing urban environments. Potholes present significant risks to vehicles, commuters, and contribute to long-term degradation of roadways. Traditional pothole detection methods are manual, time-consuming, and inefficient. This research paper proposes an automatic pothole detection system using image processing and machine learning techniques. The system captures road images via a mobile device or camera-mounted vehicle, processes them using OpenCV and convolutional neural networks (CNNs), and classifies damaged areas. The model is trained to distinguish between potholes and other road   artifacts, enabling accurate, real-time identification. Google Firebase is used to store detected locations and update road condition reports, ensuring timely maintenance action. Potholes on road surfaces pose significant challenges to transportation safety and infrastructure durability. Traditional detection methods are often manual, slow, and labor-intensive, leading to delays in maintenance and increased risks to commuters. This paper proposes an automated pothole detection system that utilizes computer vision and deep learning techniques for real-time monitoring of road conditions.

Keywords: Pothole Detection, Image Processing, CNN, OpenCV, Road Surface Monitoring, Firebase.