Camera System (ZED)

This page focuses on the hardware role of the ZED camera in field operation.

It is not a deep software API guide.

Camera Role In The Platform

The camera layer supports autonomy behavior by providing:

  • Visual scene input
  • Depth information for obstacle awareness
  • Tracking support for follow/row behaviors

If camera quality drops, autonomy behavior quality drops.

Physical Checks Before Operation

  1. Lens surfaces are clean.
  2. Camera mount is tight and aligned forward.
  3. Cable is fully seated and strain-relieved.
  4. No visible damage to housing or connector.

Environmental Readiness

For reliable behavior:

  • Use consistent, moderate lighting.
  • Avoid strong backlighting directly into lens.
  • Reduce reflective clutter near travel path when possible.
  • Keep likely targets unobstructed.

Startup Expectations

During normal boot:

  1. Compute layer comes online.
  2. Camera is detected by the stack.
  3. Autonomy modes become available once system is ready.

If modes fail to engage, include camera verification in troubleshooting.

Field Symptoms Of Camera Issues

Common signs:

  • Robot repeatedly fails to lock target in follow modes
  • Robot frequently drops target with no obvious reason
  • Row mode pauses often in clear path conditions
  • Behavior is unstable after lighting/environment changes

Operator Recovery Steps

  1. Return robot to safe idle state.
  2. Re-check lighting and clear visual clutter.
  3. Confirm camera is clean and physically aligned.
  4. Restart the autonomy workflow from idle.

If symptoms persist, escalate to support workflow for hardware inspection.

Camera Safety Notes

  • Never assume perception is perfect in all lighting.
  • Keep physical E-stop within reach during testing.
  • Validate behavior in a clear area before crowded use.