StreetSense is a Flutter-based mobile application designed to capture images of road surfaces and analyze them for damage such as potholes and cracks. Using a powerful cloud-based detection API, the app provides detailed analysis and risk assessment metrics to help users identify and evaluate road damage.
Click on the image above to watch the demo video
- Intuitive Camera Interface: Capture road images with a user-friendly camera UI
- Gallery Integration: Select existing images from device gallery
- Real-time Damage Detection: Analyze road images using cloud-based AI detection
- Visual Results: View detected damage with highlighted bounding boxes
- Damage Classification: Differentiate between potholes and cracks
- Risk Assessment Metrics:
- Overall Damage Percentage
- Hazard Level
- Accident Risk
- Comprehensive Risk Assessment
- Damage Distribution: Visual representation of damage types
- Contact Form: Submit user information for follow-up
- Responsive Design: Works across different device sizes
-
App Entry Point (
main.dart)- Initializes the Flutter environment
- Sets up the available camera(s)
- Runs the app with the CameraScreen as the home screen
-
Camera Module (
camera_screen.dart)- Provides live camera preview
- Allows image capture or gallery selection
- Includes UI controls (flash toggle, mode selection)
- Handles navigation to the Display Picture Screen
-
Image Analysis Module (
display_picture_screen.dart)- Displays the captured/selected image
- Processes the image (resizing, compression, encoding)
- Sends the image to the detection API
- Handles API response and navigation to analysis screen
-
Damage Analysis and Display Module (
damage_analysis_screen.dart)- Shows the image with overlaid detection boxes
- Calculates damage metrics
- Displays risk assessment and damage distribution
- Includes contact information form
- App starts and initializes the camera
- User captures an image or selects from gallery
- Image is processed and sent to the cloud detection API
- API returns detection results
- Results are displayed with visual overlays and metrics
- User can submit contact information or take another image
The app provides comprehensive damage analysis through multiple metrics:
- Damage Overview: A pie chart showing the overall damage percentage
- Hazard Level: Calculated based on damage type, size, and confidence
- Accident Risk: Evaluates potential for accidents based on damage characteristics
- Risk Assessment: A weighted combination of hazard level and accident risk
- Damage Distribution: Visual meters showing potholes vs. cracks distribution
- Flutter SDK (^3.5.4)
- Dart SDK
- Android Studio or Visual Studio Code with Flutter extensions
- Android SDK or iOS development setup
-
Clone the repository:
git clone https://github.com/Kanishk1420/StreetSense-Road-Damage-Detection.git -
Navigate to the project directory:
cd streetsense -
Install dependencies:
flutter pub get -
Run the app:
flutter run
The app relies on the following key dependencies:
| Package | Version | Purpose |
|---|---|---|
| camera | 0.10.5+9 | Camera access and control |
| image_picker | 1.0.7 | Accessing the device gallery |
| http | 1.1.0 | Making API requests |
| image | 4.0.15 | Image processing |
| fl_chart | 0.66.0 | Rendering charts and graphs |
| path_provider | 2.1.2 | File system access |
| sensors_plus | 6.1.1 | Device sensor access |
See pubspec.yaml for a complete list of dependencies.
The app uses the Roboflow API for road damage detection:
- API Endpoint:
https://detect.roboflow.com/road-damage-detection-apxtk/5 - Model: Road damage detection model
- Detection Classes: Potholes, Cracks (Longitudinal and Alligator)
- Confidence Threshold: Dynamically adjusted based on image and damage area
- Overlap Parameter: 20%
To use your own API key:
- Replace the API_KEY constant in
display_picture_screen.dart - Ensure your endpoint URL is correct for your model
- Code Refactoring: Breaking down
damage_analysis_screen.dartinto multiple smaller files for better maintainability - iOS Support: Adding iOS development setup and compatibility
- Performance Optimizations: Improving app response time and reducing resource usage
- Starting Screen: Creating a dedicated intro screen with animations for the StreetSense logo
- Text Animations: Adding shine/glow animations to the StreetSense text
- Button Improvements: Dynamic sizing for selected bottom buttons
- Warning Text: Enhancing visibility of warning text in the Contact Information section
- App Logo: Updating the StreetSense app logo
- Button Text: Changing "Submit" button to "Next" for better user flow
- Dark/Light Theme: Adding theme options
- Accessibility: Improving app accessibility features
- Animated Transitions: Smoother transitions between screens
- Google Maps Integration:
- Adding location tracking and mapping features
- Creating heatmaps of problematic areas
- Plotting damage locations on interactive maps
- Enhanced Registration System:
- Improved user registration page with profile management
- Historical analysis tracking
- Model Improvements:
- Higher accuracy detection for various types of potholes
- Better classification of different crack patterns (longitudinal, alligator, etc.)
- Improved detection in various lighting and weather conditions
- Offline Detection: Enabling damage analysis without internet connection
- Historical Analysis: Tracking and analyzing road damage over time
- Reporting System: Direct reporting to local authorities
- Severity Classification: Enhanced damage severity metrics
- Video Analysis: Support for analyzing video footage
- Damage Tracking: Monitoring progression of damage over time
-
Kanishk - Lead Developer
- Email: kanishkgupta2003@outlook.com
-
Kingshuk Chatterjee - UI/UX Designer
-
Md Azlan - Idea Interpreter
- Email: p038434azlannbpjc2@gmail.com
This project is licensed under the MIT License.
Developed with ❤️ using Flutter