Deep Learning | YOLOv11 | Streamlit
An AI-powered Road Damage Detection System that automatically detects and localizes different types of road damage from images using the RDD2022 Dataset and YOLOv11.
Road damage inspection is traditionally performed manually, making the process time-consuming and difficult to scale.
This project presents an AI-powered Computer Vision application that automatically detects road damage from images using a trained YOLOv11 object detection model.
The application provides an interactive Streamlit interface where users can upload road images and view detected damage with bounding boxes and confidence scores.
- Road Damage Detection
- Crack Detection
- Pothole Detection
- Detection of Multiple Road Damage Categories
- Bounding Box Visualization
- Confidence Score Analysis
- Image Upload
- YOLOv11 Object Detection
- Interactive Streamlit Dashboard
- Analytics Dashboard
- Detection History
- Annotated Detection Results
- User-Friendly Interface
| Class | Description |
|---|---|
| D00 | Longitudinal Crack |
| D10 | Transverse Crack |
| D20 | Alligator Crack |
| D40 | Pothole |
Road Image
|
v
Image Upload
|
v
Image Preprocessing
|
v
YOLOv11 Detection Model
|
v
Damage Detection
|
v
Bounding Boxes + Confidence Scores
|
v
Detection Results
- Python
- YOLOv11
- PyTorch
- Ultralytics
- OpenCV
- Pillow
- NumPy
- Pandas
- Matplotlib
- Streamlit
- Git
- GitHub
- VS Code
- Jupyter Notebook
road-damage-ai-detection/
|
├── frontend/
│ ├── app.py
│ ├── assets/
│ │ └── logo.png
│ └── ...
│
├── notebooks/
│ └── ...
│
├── requirements.txt
├── packages.txt
├── README.md
├── .gitignore
└── LICENSE
The project is based on the Road Damage Detection 2022 (RDD2022) dataset.
The dataset contains annotated road images collected from multiple countries and includes different types of road surface damage.
The model is trained to detect the following road damage categories:
- D00 - Longitudinal Crack
- D10 - Transverse Crack
- D20 - Alligator Crack
- D40 - Pothole
git clone https://github.com/WardaAhad/road-damage-ai-detection.git
cd road-damage-ai-detectionpip install -r requirements.txtstreamlit run frontend/app.pyThe application will be available at:
http://localhost:8501
Provides an overview of the Road Damage Detection System and project information.
Upload a road image and use the trained YOLOv11 model to detect road damage.
Provides visual insights and statistics related to detection results.
Allows users to review previously processed detection results.
Add Dashboard Screenshot here.
Add Detection Screenshot here.
Add Analytics Screenshot here.
Add History Screenshot here.
Try the deployed Streamlit application:
https://road-damage-ai-detection-zlum5wd8ojrh6toetrnlqb.streamlit.app/
- Automate road damage detection using Computer Vision.
- Apply YOLOv11 object detection to real-world road images.
- Detect and localize multiple types of road damage.
- Build an interactive AI application using Streamlit.
- Visualize detection results using bounding boxes and confidence scores.
- Gain practical experience in Deep Learning and Computer Vision.
- Video Road Damage Detection
- Live Camera Detection
- GPS Integration
- Automated PDF Reports
- Damage Severity Analysis
- Interactive Road Damage Maps
- Mobile Application
- Cloud-Based Storage
- Real-Time Monitoring
Bachelor of Artificial Intelligence
The Islamia University of Bahawalpur
GitHub:
If you find this project useful, consider giving the repository a star on GitHub.
Your support helps encourage further development and learning.
This project is licensed under the MIT License.
Built with Python, YOLOv11, and Streamlit.