Registration of synthetic vision imagery: issues and pitfalls [aircraft displays]

Synthetic vision (SV) and enhanced vision (EV) systems generate imagery based on databases of known items, such as terrain, runways, roads, and buildings. The computer graphics eyepoint traverses the world based on incoming navigation sensor data. The reference models and coordinate systems used for this data affect the fidelity with which the synthetic world matches the real world. Minimizing these registration errors is essential for useable SV/EV products. SV/EV systems are contrasted with simulations and games in their fundamental characteristics. Topics related to image registration accuracy are addressed for a variety of head-down, head-up, and head-mounted displays. Navigation sensors and database sources for SV/EV applications are then discussed, with an emphasis on coordinate system considerations. Finally, various earth surface models including flat, sphere, ellipsoid and geoid are discussed along with advantages and disadvantages of each. Consequences of these design choices on overall data integrity are highlighted throughout the paper.