The motion of object or camera during the duration of a television field results in a displacement of an image. Quick moving objects in each frame are blurred. The real camera is the source of blurring also. Blurring destroys contour image.

The cross correlation algorithm gives the best results in television tracking systems in difficult conditions of supervision including the motion-blurred input images. One of the purposes of the research was the experimental check of method stability with respect to motion-blurred input images.

Results of computer modeling have shown satisfactory results of recognition of the moving object parameters. The outcomes are sufficient for use. Practical results can be lower because of influence of adverse conditions of observation, appearance of coordinate distortions and influence of other errors.

In the absence of blurring, the accuracy of angular determination for angles of pitch, yaw and roll is not worse than 1 degree, and the distance to the object not less than 100 meters. Accuracy does not decrease, if the value of blurring is less than 3,5% of the linear size of moving object. When the value of blurring achieves 7%, the accuracy falls and it becomes worse than 1 degree. The model provides accuracy of coordinate measurements equal to 0,1 meters (in the absence of input noise and noise filtering).

The insensitivity of the method is not completely unexpected. Correlation on the basis of the contour description does not use all the energy of a useful signal. The motion-blurred input images destroy the contour of object. As a result, the accuracy of a method of characteristic points is lower. The cross correlation algorithm of a full image with the measurement standard ensures maximal exactness and reliability.