Computer simulation of image degradations by atmospheric turbulence for horizontal views

The development and implementation of a computer model to simulate the impact of atmospheric turbulence on image quality for a passive imaging system is presented. The presented model is an empirical one based upon the analysis of imaging distortions in real image sequences recorded under different atmospheric conditions. Only horizontal views are considered, which are typical for a ground to ground application. The computer simulation uses pristine, single images (showing no turbulence effects) as input and produces image sequences that are degraded by the specified turbulence. The implemented method can be applied for instance to the images calculated by any existing imaging simulation tool of a passive camera in a post processing step. Imagers with high frame rates can be simulated. The simulation results for a medium and a strong turbulent condition are compared to field data collected by Germany during the NATO-RTG40 White Sands Missile Range field trials of November 2005. An important feature of the presented simulation method is the consideration of the range information, which is the viewing distance to an object, or in other words, the length of the optical propagation path. In contrast to the usual way how turbulence is included into imaging simulations by assuming only a single viewing distance for all parts of a scene, different range information for different image areas can be used in our simulation. Such spatially high resolved range information can be for instance easily calculated for synthetic scenarios by computer graphics tools. Examples are presented showing the advantages of the range dependent turbulence simulation. The presented simulation method is fast in terms of computing time and well suited for real-time simulations using the computing power of nowadays graphics processors.