This project aims to detect potholes in road images using machine learning techniques. The goal is to assist in identifying and prioritizing road maintenance efforts to improve road safety and infrastructure.
Note: I created this project as a practice to apply for the Google Summer of Code (GSoC) 2024 project for the City of Boston. It is a sample project created for practicing machine learning and web development skills.
- Upload an image to detect potholes.
- Display the prediction result indicating the presence of potholes in the image.
- Python
- OpenCV (cv2) - for image processing
- NumPy - for numerical computing
- TensorFlow - for building and training the model
- os - for file operations
- Flask
- TensorFlow
- HTML/CSS
- JavaScript
Note: If you encounter an error while activating scripts in the virtual environment, run the following command:
Set-ExecutionPolicy -Scope Process -ExecutionPolicy Bypass- Upload an image containing a road surface.
- Click the "Predict" button to detect potholes in the image.
- View the prediction result, which indicates whether potholes are present in the image.
