Towards an Augmented Reality System for Obstacle Avoidance in Agricultural Machinery

We present a work-in-progress AR system designed to support obstacle avoidance in agricultural vehicles. Static hazards such as wells, rocks, or ditches are often obscured by tall crops and lack consistent physical markers, posing a risk to large machinery during field operations. Our prototype uses a handheld tablet and RTK-GNSS to display virtual warning markers that mimic real-world signage. These markers are color-coded and icon-based, conveying obstacle type and urgency with minimal cognitive load. The system segments the field into priority zones based on vehicle dimensions and movement direction. A preliminary field test shows that the AR interface improves hazard awareness even for fully occluded objects. Future work will address user studies, dynamic obstacles, and integration into vehicle-mounted displays.