Visual navigation systems for drones

Most drones navigate with GNSS and inertial sensors. When GNSS is jammed, spoofed, blocked by buildings or simply unavailable, the inertial solution drifts within minutes. Visual navigation systems use cameras, and often other sensors, to estimate position and velocity without satellites.

Techniques

• Visual odometry and visual-inertial odometry (VIO): track features between camera frames and fuse them with the IMU to estimate motion. Drift grows with distance but much more slowly than with inertial sensors alone.
• Terrain and map matching: compare camera images with stored maps or satellite imagery to fix absolute position.
• Visual landing aids: recognise markers, runways or landing pads for precision landing.
• Optical flow sensors: downward-looking sensors that measure velocity over the ground, common for hover hold at low altitude.

Key specifications

• Position accuracy and drift over a given distance or time without GNSS.
• Operating envelope: altitude, speed, light conditions and terrain texture. Water, snow, fog and night are difficult for visual systems.
• Sensors: mono or stereo cameras, thermal cameras for night, laser rangefinders.
• Processing: integrated computer or software for an existing mission computer.
• Integration: outputs to the flight controller as position, velocity or GNSS-like messages.
• Size, weight and power.

Why it matters

GNSS interference has become common in conflict areas and near sensitive sites, and urban eVTOL operations need navigation that does not depend on satellites alone. Visual navigation complements robust GNSS receivers and high-grade IMUs.

Manufacturers in the directory

Examples include Honeywell Aerospace, Auterion and ModalAI.