- Camera falsely detects obstacles when dirty or smudged.
- Increased false detections in poor lighting conditions.
- Pathways sometimes mis-detect obstacles as parked vehicles.
- Dirty or obstructed camera lenses.
- Reduced camera accuracy in low-light environments.
- Current depth estimation model not fully optimized for all scenarios.
- Clean the camera lens thoroughly and check again.
- Perform a Camera Calibration to reset the baseline settings.
- Use “Vision Zone” to cover areas you know are clear of obstacles or where obstacle avoidance is unnecessary.
- Adjust the Obstacle Avoidance Mode based on the cause:
- Camera frequently dirty/blocked→ Keep Moderate Bypass: If the camera becomes unavailable, Yarbo will continue mowing using the bumper for obstacle detection (plus ultrasonic, radar, if equipped). Once vision is available again, Moderate Bypass resumes normal vision-based obstacle avoidance, balancing mowing continuity and safety.
- Other persistent false detections → Try Gentle Contact temporarily, but only in areas clear of children, pets, valuable objects, or other items that could be damaged by contact.
- Issue limited to one location → Use a Vision Zone to apply Gentle Contact only where needed while keeping more protective settings in sensitive areas.
- Continuous optimization of the depth estimation model and additional training on a larger dataset.
- Firmware update planned for future releases to improve detection accuracy.