108 lines
3.6 KiB
Markdown
108 lines
3.6 KiB
Markdown
# How to Use Smart Intersection Features in Desktop App
|
|
|
|
This guide explains how to use the Smart Intersection analytics features integrated into the Traffic Monitoring Desktop Application.
|
|
|
|
## Overview
|
|
|
|
The Smart Intersection features provide advanced scene-based analytics for traffic monitoring, including multi-camera object tracking and region-of-interest analysis.
|
|
|
|
## Getting Started
|
|
|
|
### 1. Launch the Application
|
|
- Run `python main.py` from the qt_app_pyside1 directory
|
|
- The application will start with Intel Arc GPU acceleration enabled
|
|
- Wait for the splash screen to complete initialization
|
|
|
|
### 2. Access Smart Intersection Features
|
|
|
|
#### Configuration Panel
|
|
1. Navigate to the **Config** tab in the main window
|
|
2. Select **Smart Intersection** from the configuration categories
|
|
3. Configure the following settings:
|
|
- **Camera Settings**: Add multiple camera sources for scene analytics
|
|
- **Tracker Parameters**: Adjust tracking sensitivity and frame rate handling
|
|
- **ROI Definition**: Define regions of interest for analytics
|
|
- **Performance Settings**: Optimize for your Intel Arc GPU
|
|
|
|
#### Analytics Dashboard
|
|
1. Go to the **Analytics** tab
|
|
2. View real-time scene analytics including:
|
|
- Multi-camera object tracking
|
|
- Region-based event detection
|
|
- Traffic flow analysis
|
|
- Pedestrian safety monitoring
|
|
|
|
#### VLM Insights
|
|
1. The VLM insights panel provides AI-powered analysis
|
|
2. Enable scene-based insights for enhanced understanding
|
|
3. View contextual information about detected events
|
|
|
|
## Key Features
|
|
|
|
### Multi-Camera Scene Analytics
|
|
- **Object Tracking**: Track objects across multiple camera views
|
|
- **Scene Fusion**: Combine data from multiple cameras for comprehensive view
|
|
- **Real-time Processing**: All processing occurs locally with GPU acceleration
|
|
|
|
### Region of Interest (ROI) Analytics
|
|
- **Crosswalk Monitoring**: Detect pedestrians in crosswalk areas
|
|
- **Lane Analysis**: Monitor vehicle dwell time and traffic flow
|
|
- **Safety Zones**: Define areas for safety monitoring and alerts
|
|
|
|
### Performance Optimization
|
|
- **Intel Arc GPU**: Optimized for Intel Arc GPU acceleration
|
|
- **Local Processing**: No cloud dependency, all processing local
|
|
- **Real-time Feedback**: Immediate response to traffic events
|
|
|
|
## Configuration Options
|
|
|
|
### Tracker Settings
|
|
```json
|
|
{
|
|
"max_unreliable_frames": 10,
|
|
"non_measurement_frames_dynamic": 8,
|
|
"non_measurement_frames_static": 16,
|
|
"baseline_frame_rate": 30
|
|
}
|
|
```
|
|
|
|
### Camera Configuration
|
|
- **Camera ID**: Unique identifier for each camera
|
|
- **Position**: Physical position in the intersection
|
|
- **Calibration**: Camera calibration parameters
|
|
- **ROI Mapping**: Define which regions each camera monitors
|
|
|
|
## Troubleshooting
|
|
|
|
### Common Issues
|
|
1. **GPU Not Detected**: Ensure Intel Arc GPU drivers are installed
|
|
2. **Poor Tracking**: Adjust tracker parameters in configuration
|
|
3. **Performance Issues**: Check GPU utilization in Performance tab
|
|
|
|
### Performance Tips
|
|
- Use appropriate frame rates for your hardware
|
|
- Configure ROI regions efficiently
|
|
- Monitor GPU memory usage
|
|
- Adjust tracking parameters based on scene complexity
|
|
|
|
## Advanced Usage
|
|
|
|
### Custom ROI Definition
|
|
1. Load video or connect cameras
|
|
2. Pause on a representative frame
|
|
3. Use the ROI drawing tools to define areas
|
|
4. Save configuration for reuse
|
|
|
|
### Integration with Existing Workflows
|
|
- Export analytics data for further analysis
|
|
- Configure alerts for specific events
|
|
- Integrate with external systems via configuration
|
|
|
|
## Support
|
|
|
|
For additional help:
|
|
- Check the **Help** → **User Guide** menu
|
|
- Review system requirements
|
|
- Consult troubleshooting documentation
|
|
- Check performance metrics in the Performance tab
|