Smart Lane Track Pothole Detection and Tracking System

Project Overview

This project aims to automate the detection and tracking of potholes on roads using aerial drone footage. Leveraging a custom-trained YOLO12 model, the system processes live or recorded video feeds to identify potholes, mark their positions, and assign them unique IDs for temporal tracking. This is especially useful for large-scale road infrastructure monitoring and maintenance planning.

Tech Stack

Key Features

• Drone-Based Input
  Efficient pothole detection from top-down views.

• Custom YOLO12 Model
  Trained specifically to detect potholes in road images.

• Pothole Tracking with IDs
  Each pothole is assigned a persistent ID, helping track it across frames.

• Real-Time Performance
  Processes every third frame to balance speed and accuracy.

• Visual Output
  Overlays bounding boxes, labels, and IDs on the video feed.

• Error Handling
  Detects if the video cannot be opened and exits gracefully.

Project Workflow

1. Model Loading

   • YOLO12 model (best.pt) is loaded for inference.

2. Video Processing

   • Reads video frames from a drone footage file e.g: (dronepotholes.mp4).

   • User can also use live camera

3. Pothole Detection & Tracking

   • YOLO detects potholes and assigns tracking IDs using model.track().

4. Visualization

   • Bounding boxes and pothole IDs are drawn on each processed frame.

5. Exit Conditions

   • Stops when video ends or when user presses Esc.

Use Cases

• Smart City Road Maintenance
  Automate pothole reporting and send location-based maintenance alerts.

• Government Infrastructure Audits
  Use aerial footage to assess road quality in large geographic areas.

• Web-based Road Health Dashboards
  Integrate this system with dashboards that display road damage status.

• Autonomous Vehicle Navigation
  Aid self-driving cars by feeding pothole maps into path planning systems.

Conclusion

SmartLaneTrack transforms drone footage into actionable insights for infrastructure maintenance using state-of-the-art computer vision. It eliminates the need for manual inspection by automatically identifying, labeling, and tracking potholes with high accuracy and speed using YOLO12.

This project is a scalable solution for governments, municipalities, and transport departments to reduce road accidents, optimize maintenance budgets, and enhance road safety.