cd ~/
CalibrationComputer VisionManufacturingQualityMagic Leap

Camera Calibration: Theory Meets Manufacturing Reality

The gap between textbook calibration and production-scale calibration of perception systems - where precision meets throughput.

Evyatar Bluzer
2 min read

Calibration is the unsexy foundation of every perception system. Get it wrong, and nothing else matters.

What Needs Calibrating

Intrinsic parameters: Focal length, principal point, distortion coefficients for each camera.

Extrinsic parameters: Relative pose between every sensor pair (camera-camera, camera-IMU, camera-depth).

Temporal alignment: Latency offsets between sensors.

Depth-specific: ToF phase offsets, systematic depth errors, multi-path correction tables.

The Textbook Approach

Standard calibration: wave a checkerboard in front of the device, detect corners, run bundle adjustment.

Works great in a lab with one prototype and an expert operator.

Fails completely at manufacturing scale with:

  • 10 seconds per device
  • Operators with minimal training
  • Environmental variation (temperature, lighting)
  • Thousands of devices per day

Production Calibration Architecture

We're designing a calibration cell:

┌────────────────────────────────────────────┐
│              Calibration Cell              │
│  ┌────────────────────────────────────┐   │
│  │      Controlled Environment        │   │
│  │  - Temperature: 25°C ± 1°C        │   │
│  │  - Humidity: 45% ± 5%             │   │
│  │  - Lighting: 500 lux diffuse      │   │
│  └────────────────────────────────────┘   │
│                                            │
│  ┌──────────────┐    ┌──────────────┐    │
│  │  Reference   │    │    Device    │    │
│  │   Targets    │    │   Under      │    │
│  │  (3D known   │    │   Test       │    │
│  │   geometry)  │    │              │    │
│  └──────────────┘    └──────────────┘    │
│                                            │
│  ┌────────────────────────────────────┐   │
│  │      Automated Motion System       │   │
│  └────────────────────────────────────┘   │
└────────────────────────────────────────────┘

The motion system moves reference targets through the field of view, collecting data across the full calibration manifold automatically.

Key Insights

Temperature matters: Optical systems change with temperature. A camera calibrated at 25°C may be off at 35°C operating temperature. We need thermal models.

Calibration lifetime: Parameters drift over time (mechanical settling, thermal cycling). How often do we need to recalibrate?

Per-unit vs. golden: Some parameters can use nominal values; others must be measured per-unit. Finding this boundary is crucial for throughput.

Self-calibration: Can the device recalibrate itself in the field? Essential for longevity.

Accuracy Budgets

Working backwards from user experience:

  • Virtual object placement: ±1cm accuracy
  • Requires 6DoF tracking: ±1mm, ±0.1°
  • Requires intrinsic calibration: under 0.2 pixel error
  • Requires extrinsic calibration: under 0.1° rotation, under 0.5mm translation

These are tight tolerances for a consumer device.

More on depth calibration specifically next month.

Comments