Research On IR Imaging Tracking Algorithm For Targets On The Ground

The infrared imaging tracking algorithm has the definitive effect on theperformance of the infrared imaging guidance system. First, the correlation matchingalgorithm which is often used to track targets on the ground is studied in thisdissertation, from the aspects of similarity measurement, partitioning and switching ofthe states of the tracking system, systematic error and so on. Because thecharacteristic that the template changing method influences the stability of thetracking system directly, we design a template changing method based on Kalmanfilter in this dissertation. It can change the template properly when the target changesits figure or size, which guarantees the correction of the matching result. At last, theanalysis of some real sequence images shows that the algorithm is feasible and valid.The linear prediction and the square prediction algorithm are simulated in theother part of this dissertation. We obtain the law that the predicting precision variedwith the number of the predicted value, then we presents a kind of integrativeprediction algorithm based on the two prediction algorithm. The simulation resultsshow that this algorithm is better than the two prediction algorithm when they areused separately.