A maximum likelihood estimator for tracking purposes with extended sources

In many cases, optical tracking systems do not have cooperative beacons available. This is particularly true for the case involving tracking the reflectance from a laser illuminated target such as a missile seeker head, where the object of interest is most definitely an extended source. Furthermore, the extended source is always combined with noise such as shot noise which further degrades the signal. Consideration is also given to atmospheric turbulence. This paper examines the performance of an existing projection-based, maximum-likelihood technique for tilt estimation in the presence of extended sources, with particular application to the image motion tracking problem. Comparison is made between the performance of a traditional centroiding algorithm and a projection-based algorithm with simulated data. The projection-based algorithm is shown to offer improved performance in the motion tracking problem.