Adrenaline Icon

Setting Up Cameras

Cameras are critical for AI-powered timing and scoring. Set up cameras at the finish line and checkpoints to enable automated rider detection and lap timing. TrackLang defines camera position, orientation, and field of view—making accurate camera configuration essential for reliable AI Vision processing.

Why Use Cameras?

Cameras provide visual timing data that Adrenaline's AI Vision system processes to detect riders, track positions, and record crossing times. When properly configured, cameras enable fully automated timing without manual intervention.

Key Benefits

  • Automated timing—no manual lap recording needed
  • AI Vision detects riders crossing finish line and checkpoints
  • Multiple cameras provide coverage of entire track
  • Video evidence for disputed results or close finishes
  • Works in conjunction with automated timing for redundant verification
  • Enables post-race analysis and highlights

Camera Use Cases

Finish Line Camera

The primary camera. Positioned perpendicular to traffic flow at the finish line. Captures riders as they complete laps. Essential for accurate lap timing.

Checkpoint Cameras

Additional cameras at key points around the track (split timing zones, turns, whoops sections). Useful for sector times, race monitoring, and safety.

Broadcast/Media Cameras

Cameras placed for spectator viewing or broadcast. Not necessarily for timing, but can provide additional angles for analysis.

Safety Cameras

Positioned at high-risk sections to monitor crashes or hazards. May not be used for timing but valuable for race control.

Physical Camera Placement

Finish Line Camera: Best Practices

  • Perpendicular angle: Position camera perpendicular (90°) to the flow of traffic. This gives the clearest view of riders crossing the line.
  • Viewing angle: Position at horizon level, or up to 60° downward angle. Avoid extreme downward angles (>60°) which distort rider detection.
  • Distance: Far enough to capture the full width of the track, close enough for clear rider identification. Typically 10-30 feet from finish line.
  • Height: Elevated view (6-15 feet high) works best. Too low = riders obscured. Too high = extreme angle issues.
  • Stability: Mount securely to prevent camera shake from wind, vibration, or crowd movement.
  • Clear sightline: No obstructions (fences, banners, people) between camera and finish line.

Camera Field of View

Ensure the finish line is fully visible in frame:

  • Full track width visible—riders can't cross outside camera frame
  • Include 5-10 feet before and after finish line for context
  • Avoid zooming too far in—need buffer space for wide starts or drift
  • Test by having someone ride through—verify they're visible throughout crossing

Lighting Considerations

Good lighting is critical for AI detection:

  • Avoid backlighting: Don't point camera directly at the sun—riders become silhouettes
  • Consistent exposure: AI works best with stable brightness. Avoid rapid shadows or flickering
  • Night racing: Requires floodlights at finish line for camera visibility
  • High contrast: Riders should be clearly distinguishable from background

Multiple Camera Setup

Multiple cameras are supported and encouraged. Use additional cameras for checkpoints, sector timing, and redundancy.

Benefits of Multiple Cameras

  • Redundancy: If one camera fails or has obstructed view, others provide backup
  • Sector timing: Multiple checkpoints enable split times throughout the lap
  • Race monitoring: See multiple parts of track simultaneously for better race control
  • Improved accuracy: Cross-reference data from multiple cameras for higher confidence
  • Comprehensive coverage: Catch riders even if they take unusual lines or drift wide

Best practice: Start with 1 finish line camera. Once working, add checkpoint cameras at key locations (first turn, big jump, halfway point, etc.).

TrackLang: Camera Configuration

TrackLang is critical for camera setup. It defines the camera's position, orientation, and field of view in relation to the track. AI Vision uses this information to correctly process video and detect riders.

Why TrackLang Matters

Without accurate TrackLang camera configuration, AI Vision cannot properly interpret video. The system needs to know where the camera is, which direction it's pointing, and what area of track it covers. Incorrect TrackLang = missed detections or false positives.

TrackLang Camera Parameters

When defining cameras in TrackLang, specify:

  • Location: Exact distance from start line (pos/neg) of camera position
  • Offset: Exact distance away from the track (perpendicular to) of camera position
  • Orientation (heading): Direction camera is pointing (0-360°, where 0 = North)
  • Tilt angle: Vertical angle (0° = horizon, positive = downward tilt, up to 60°)
  • Field of view: Horizontal angle coverage (usually 50-120° depending on lens)
  • Coverage zone: Which timing zone/checkpoint the camera monitors (finish line, checkpoint 1, etc.)
  • Height above ground: Camera elevation in feet/meters

Example TrackLang Camera Definition

camera finish_line {
  position: 34.123456, -118.654321
  heading: 270  // pointing West
  tilt: 30      // 30° downward
  fov: 90       // 90° horizontal field of view
  height: 10    // 10 feet above ground
  zone: "finish"
}

This tells AI Vision exactly how to interpret video from this camera.

Determining Camera Parameters

How to find TrackLang values:

  • Record rough distance from the start line to the camera position
  • Record rough distance away from the track (perpendicular to) of camera position
  • Heading: Use compass app or visual estimate (North=0°, East=90°, South=180°, West=270°)
  • Tilt: Measure with inclinometer app or estimate visually
  • FOV: Check camera/lens specs. Typical action cameras: 90-120°. Zoom lenses: 30-60°.
  • Height: Measure from ground to camera lens

Camera Hardware Recommendations

Resolution

Minimum 1080p (1920x1080). 4K preferred for best AI detection. Higher resolution = better accuracy, especially at distance.

Frame Rate

30fps minimum. 60fps preferred for fast-moving riders. Higher frame rates reduce motion blur and improve detection.

Lens/Field of View

Wide angle (90-120°) for finish line coverage. Can use narrower FOV for checkpoints. Avoid fisheye lenses (distortion confuses AI).

Weatherproofing

Outdoor races require weatherproof cameras (IP65+ rating). Protect from rain, dust, and mud spray.

Power Supply

Continuous power recommended (AC adapter or large battery bank). GoPro-style battery swapping works but risks missing races.

Connectivity

WiFi or wired connection to streaming device. Live streaming to Adrenaline servers enables real-time AI processing.

Camera Setup Process

Step 1: Physical Installation

  1. Choose camera location (finish line, checkpoint, etc.)
  2. Mount camera securely—tripod, fence mount, scaffold, etc.
  3. Adjust camera angle to perpendicular viewing position
  4. Set tilt to horizon or up to 60° downward
  5. Verify entire finish line/zone is visible in frame

Step 2: Measure Camera Parameters

  1. Record rough distance from the start line to the camera position
  2. Record rough distance away from the track (perpendicular to) of camera position
  3. Measure heading/orientation (compass app)
  4. Measure tilt angle (inclinometer app or visual estimate)
  5. Note field of view from camera/lens specs
  6. Measure height above ground

Step 3: Define Camera in TrackLang

  1. Open Track Developer and edit track's TrackLang code
  2. Add camera definition with measured parameters
  3. Assign camera to timing zone (finish, checkpoint1, etc.)
  4. Save TrackLang changes

Step 4: Connect Camera to Adrenaline

  1. Start camera streaming (WiFi, RTMP, or other supported protocol)
  2. Configure Adrenaline to receive camera feed
  3. Verify live video appears in Race Control interface
  4. Test AI detection with a rider crossing the zone

Step 5: Calibrate & Test

  1. Run test crossings with known riders
  2. Verify AI detects crossings accurately
  3. Adjust TrackLang parameters if detections are off
  4. Fine-tune camera angle or position as needed
  5. Confirm multiple test runs before using in live races

Troubleshooting Common Issues

AI not detecting riders

Possible causes:

  • Incorrect TrackLang camera parameters (position, heading, tilt, FOV)
  • Poor lighting—riders too dark or backlit
  • Camera angle too extreme (>60° tilt)
  • Low resolution or frame rate
  • Camera feed not connected to Adrenaline

False detections (detecting riders when none present)

Possible causes:

  • Moving objects in frame (spectators, flags, animals)
  • Camera shake or vibration
  • TrackLang zone definition too wide
  • Need to adjust AI sensitivity settings

Camera feed quality poor

Possible causes:

  • Low bandwidth—compress stream less or upgrade connection
  • Camera dirty/foggy lens—clean regularly
  • Backlighting causing exposure issues—reposition camera
  • Motion blur from fast riders—increase frame rate

TrackLang parameters seem wrong

Solutions:

  • Re-measure GPS coordinates with better accuracy
  • Use compass app to verify heading
  • Check camera specs for actual FOV (not marketing claims)
  • Verify tilt angle with inclinometer
  • Test with known rider crossing and adjust iteratively

Best Practices

  • Start simple: One finish line camera first. Add checkpoints once primary camera works reliably.
  • Test early: Set up cameras days before the event to allow testing and adjustment time.
  • Document everything: Record TrackLang parameters, camera positions, and settings for future events.
  • Backup cameras: If possible, have a second finish line camera from a different angle for redundancy.
  • Stable mounts: Invest in quality mounts. Wobbly cameras = unusable footage.
  • Clean lenses: Check before each race day. Mud and dust accumulate quickly at MX tracks.
  • Power management: Always have backup batteries or redundant power supplies.
  • TrackLang accuracy: Spend time getting TrackLang parameters correct. This is the foundation of AI detection.

Common Questions

What cameras are compatible with Adrenaline?

Most cameras that can stream via WiFi, RTMP, or similar protocols. GoPros, action cameras, webcams, and IP cameras all work if they can provide a live feed.

Can I use recorded video instead of live streaming?

Post-race video processing is possible but not real-time. For live timing during races, you need live camera streaming.

How accurate is AI Vision timing?

With proper camera setup and TrackLang configuration, AI detection is frame-accurate (within 1/30th or 1/60th of a second depending on frame rate).

Do I need cameras at every checkpoint?

No. Cameras are optional. You can use automated timing without cameras. But cameras provide more accurate timing and enable features like automated scoring.

What if TrackLang parameters are slightly off?

Small errors (±5° heading, ±1 meter position) are usually acceptable. AI is somewhat tolerant. Large errors cause missed detections.

Can I reuse camera setups across multiple events?

Yes, if the track and camera positions are identical. TrackLang is track-specific, so same track = same TrackLang camera config.

How do I test if my camera setup is working?

Run ad-hoc races or test sessions with known riders. Have them cross the finish line multiple times and verify AI detects each crossing.

Related Topics

Tags
trackdevelopmentcamerastimingai-vision

Was this information useful?

Help us improve our documentation by rating this article and sharing your feedback.