Bayesian statistical approaches to tracking through turbulence

We describe two Bayesian approaches for tracking through turbulence. The problem considered involves an extended target actively illuminated with several lasers. The returned imagery is used to infer atmospheric tilt. The main application of this technology is the control of a steering mirror, which is used to point a laser weapon at the target. Several model-based approaches are examined, including an optical transfer function model and a tilted reference image model.