Enhancement of TV Tracking Effeciency Using Deterministic and Statistical Automatic Control Predictors

The tracking process is simply the continuous determination of the target coordinates. The tracking of moving objects in noisy conditions is very complex. The use of spatial band-pass filters (window) reduces the effect of all objects lying outside this window. The crossing object is a moving object flying across the window at the same time with intended target. So, the tracking system may be confused and become unable to recognize the correct target after separation. If the system is designed to have the ability to predict the new target center relative position, the window will be located at the predicted center, which is assumed to be very close to the real position. As a result, the system performance to track maneuvering targets will be improved. This paper introduces the use of different kinds of predictors to overcome the two main problems of tracking process (crossing moving targets and target maneuver). Simulation results show that the use of statistical predictors gives useful performance in a large class of tracking problems, but it increases the system complexity due to the need to compute the mean and variance of the target pixels inside the window.