This project implements a real-time pothole detection system using computer vision techniques integrated into an Android mobile application.
The application processes live camera frames from a smartphone and detects potholes on road surfaces. The goal is to assist in road condition monitoring and infrastructure maintenance by automatically identifying potholes during vehicle movement.
The detection pipeline combines image processing, object tracking, and machine learning techniques to detect road surface anomalies.
- Real-time pothole detection using smartphone camera
- Computer vision–based object detection
- Android mobile application interface
- Native image processing using JNI and C++
- Efficient frame processing for mobile devices
- Demo APK included for testing
- Camera Input
- Frame Capture (Android Camera API)
- YUV → RGB Conversion
- Image Processing using OpenCV / Native C++
- Object Detection & Tracking
- Pothole Identification
- Bounding Box Visualization
- Android Studio
- Java
- C++
- JNI (Java Native Interface)
- OpenCV
- Python
- Jupyter Notebook
- Download app-release.apk
- Transfer it to an Android device
- Install the APK
- Launch the application
- Allow camera access
The app will start detecting potholes in real time.
Clone the repository
git clone https://github.com/Sowdeshwar-99/real-time-pothole-detection-android.git
Open the project in Android Studio
Build and run on an Android device.
The notebook PotholeDetection.ipynb contains the experimentation and development steps for the pothole detection model including:
Image preprocessing
Feature extraction
Model experimentation
Evaluation
Applications
Road infrastructure monitoring
Smart city development
Automated road inspection systems
Driver assistance systems
Future Improvements
GPS-based pothole mapping
Cloud-based reporting system
Deep learning model integration
Integration with municipal road monitoring systems
This project was developed during a competitive Designathon, where teams were challenged to build innovative technology solutions for real-world problems.
Our team developed a real-time pothole detection system using computer vision and an Android application to assist with road infrastructure monitoring.
The project was selected as a winning solution, which led to securing an internship opportunity with Ericsson.
- Sowdeshwar Survesha Kumar
- Sudharshan Govindan
This project was developed collaboratively as part of the Designathon.
Original repository reference:
https://github.com/sudharshan1234/PotholeDetector