Abstract :

Road Infrastructure Plays A Crucial Role In Transportation Efficiency And Public Safety. However, Potholes Remain One Of The Most Common And Hazardous Issues Affecting Road Quality, Leading To Vehicle Damage, Traffic Congestion, And Accidents. Traditional Methods Of Pothole Detection Rely Heavily On Manual Inspection, Which Is Time-consuming, Inefficient, And Prone To Human Error. With Advancements In Deep Learning And Computer Vision, Automated Pothole Detection Systems Have Emerged As A Promising Solution To Address These Challenges.This Project Presents A Real-time Pothole Detection System Using Deep Learning-based Object Detection Techniques. The System Leverages The YOLO (You Only Look Once) Model, A State-of-the-art Algorithm Known For Its Speed And Accuracy In Detecting Objects Within Images And Videos. The Model Is Trained On A Dataset Containing Pothole Images And Is Capable Of Identifying Potholes In Both Static Images And Real-time Video Streams. The Integration Of ByteTrack Tracking Algorithm Enhances The System’s Ability To Track Detected Potholes Across Frames, Ensuring Continuity And Improved Detection Reliability.The Proposed System Is Implemented Using Streamlit, Providing An Interactive Web-based Interface That Allows Users To Upload Images Or Videos Or Use Sample Data For Testing. The System Processes Input Data, Detects Potholes, And Displays Annotated Results With Bounding Masks Highlighting The Detected Regions. Additionally, It Provides Real-time Performance Metrics Such As Frame Rate (FPS), Ensuring Transparency In System Performance.The Use Of Segmentation-based Detection Further Improves The Accuracy Of Pothole Identification By Precisely Outlining The Affected Areas Rather Than Just Providing Bounding Boxes. This Enables Better Assessment Of Pothole Size And Severity, Which Can Be Useful For Road Maintenance Planning. The System Also Allows Customization Of Parameters Such As Confidence Thresholds And Frame Size, Making It Adaptable To Different Use Cases And Environments.Experimental Results Demonstrate That The Proposed System Achieves High Accuracy And Efficiency In Detecting Potholes Under Various Lighting And Road Conditions. The Integration Of Deep Learning Models With Realtime Processing Capabilities Makes This System Suitable For Deployment In Smart Transportation Systems, Autonomous Vehicles, And Road Monitoring Applications.In Conclusion, This Project Highlights The Potential Of Deep Learning And Computer Vision In Automating Infrastructure Monitoring Tasks. The Proposed Pothole Detection System Not Only Reduces Manual Effort But Also Enhances Road Safety By Enabling Timely Identification And Repair Of Road Damages. Future Work May Include Integrating GPS Tagging For Detected Potholes And Deploying The System On Edge Devices For Large-scale Real-world Implementation.

Published:

07-4-2026

Issue:

Vol. 26 No. 4 (2026)

Page Nos:

1489-1497

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite