Tracking Model of Moving Target Based on KNN - SVM

According to the defects of KNN(K-Nearest Neighbor) algorithm and SVM(Support Vector Machine) algorithm in tracking a moving target such the large consumption and the low accuracy of target tracking error, a tracking model of moving target is proposed based on the combination of KNN algorithm and SVM algorithm with minimum distance optimization. First categories divided according to the principle of minimum distance classifier, to optimize the classification accuracy of standard KNN algorithm. And then the SVM as the KNN classifier with each type of only one representative point, and the improved KNN classifier is used to improve the accuracy of classification of SVM classifier boundary surface, and the SVM classifier is used to reduce the operand of the improved KNN algorithm. The simulation experiments show that the proposed improved KNN algorithm has higher target tracking accuracy and less computation, and has good effect in the actual application of tracking a moving target.