Event Detection
Understanding how NomadicML detects and analyzes driving events
Event Detection
NomadicML’s core capability is its ability to automatically detect and analyze key driving events. This page explains how event detection works and how to interpret the results.
Event Types
NomadicML identifies several categories of driving events:
Traffic Violations
Events that may constitute violations of traffic laws:
- Running stop signs or red lights
- Improper lane changes
- Failure to yield
- Speeding
- Illegal turns
Safety Alerts
Events that indicate potential safety concerns:
- Near collisions
- Hard braking
- Unsafe following distance
- Distracted driving indicators
Drive Quality
Events related to overall driving technique:
- Jerky acceleration or braking
- Poor lane positioning
- Hesitation at intersections
- Inefficient routing
DMV Compliance
Events specifically relevant to DMV testing and regulations:
- Parallel parking execution
- Three-point turn technique
- Highway merging behavior
- Proper signaling
Event Detection Process
The NomadicML platform uses a multi-stage analysis process:
- Frame Analysis: The video is processed frame-by-frame to detect objects, movements, and road features
- Motion Tracking: Objects are tracked across frames to understand their movement and relationships
- Pattern Recognition: AI models identify patterns that constitute specific driving events
- Context Understanding: The system considers the driving environment and circumstances
- Rule Application: Traffic rules and best practices are applied to identify violations and issues
Event Details
Each detected event provides comprehensive information:
Event Card
The event card shows:
- Timestamp: When the event occurred in the video
- Event Type: The category of the detected event
- Severity: Low, medium, or high impact
- Description: A concise explanation of what happened
- Thumbnail: Visual snapshot of the event
Detailed Analysis
Click on any event to view detailed analysis:
Event Information
- Time Range: Start and end time of the event
- DMV Rule: Relevant driving regulations
- Location: Where in the video the event occurred
- Objects Involved: Vehicles, pedestrians, or other elements
AI Analysis
- Context: Detailed description of the driving scenario
- Violation Details: Specific aspects of the violation or issue
- Safety Implications: Potential risks associated with the event
- Recommendations: Suggestions for improvement
Severity Levels
Events are categorized into three severity levels:
Low Severity (Blue)
- Minor technical issues
- No safety risk
- Common driving habits that could be improved
Medium Severity (Yellow)
- Clear rule violations
- Potential for safety issues
- Significant driving technique problems
High Severity (Red)
- Serious safety concerns
- Major traffic violations
- Critical driving errors
Viewing Events
You can access events in several ways:
Video Timeline
When viewing a video, events appear as markers on the timeline:
- Color-coded by severity
- Positioned at the exact point of occurrence
- Hover for quick preview
- Click to jump to that point in the video
Event List
Each video has a dedicated event list showing all detected events:
- Sortable by time, type, or severity
- Filterable to focus on specific event types
- Expandable for quick viewing of details
Consolidated Events
The Events section of the platform shows events across all videos:
- Group by video, date, or driver
- Search for specific event types
- Identify patterns across multiple drives
Working with Events
NomadicML offers several ways to interact with detected events:
Adding Notes
You can attach notes to any event:
- Click the Add Note button in the event details
- Enter your observations or feedback
- Save to attach the note to the event
Flagging Events
If you disagree with the detection:
- Click the Flag button
- Select a reason (False Positive, Severity Issue, etc.)
- Add optional comments
- Submit the flag for review
Exporting Events
Export event data for external use:
- Select the events you want to export
- Click the Export button
- Choose your preferred format (CSV, PDF, etc.)
- Download the export file
Event Statistics
NomadicML provides aggregated statistics about detected events:
- Frequency: How often each event type occurs
- Trends: Changes in event patterns over time
- Comparisons: How events compare to benchmark data
- Hotspots: Locations or situations where events commonly occur
These statistics help identify patterns and areas for improvement.
Next Steps
Now that you understand event detection, learn about: