Time-to-go estimation from infrared images

The author deals with the problem of an aircraft flying towards a rendezvous point using passive image data to calculate the flight duration. A technique is described which estimates time-to-go from a sequence of infrared images by applying a generalisation of a conventional correlation process. The method relies on measurements from a multiple point correlation tracker, which is able to track a landmark or object. It is shown that a difference squared correlator is equivalent to the measurement equation of a parameter estimation scheme with tracking offsets as states. A simple relative motion model is postulated, which uses magnification, rotation and translation, with the instantaneous measure of time-to-go related to the magnification scale factor. The mapping of tracking errors to errors in the estimated motion parameters is investigated.<>