基于AdaBoost.ECOC的合成孔径雷达图像目标识别研究

为了提高合成孔径雷达图像目标识别系统的性能,提出了一种合成孔径雷达图像目标识别的新方法,结合纠错输出码对基本AdaBoost算法进行多类别推广,并将推广后的算法(AdaBoost.ECOC)应用于合成孔径雷达图像目标识别.用运动和静止目标获取与识别数据库中的三类地面军事目标进行识别实验,并将识别结果与其他识别方法进行比较.实验结果表明,提出的基于AdaBoost.ECOC的识别算法可以有效地应用于合成孔径雷达目标识别,并能显著提高目标识别系统的识别性能.

Research on synthetic aperture radar image target recognition based on AdaBoost.ECOC

A new method for synthetic aperture radar image target recognition was proposed, which extended the basic AdaBoost algorithm for multi-class classification, and the new algorithm (AdaBoost.ECOC) was applied to synthetic aperture radar image target recognition. The extended algorithm was applied in the recognition experiment on three types of ground military vehicles in MSTAR database and the result was compared with other recognition algorithms. Results were presented to verify that the performance of the recognition system was improved significantly, and the method presented in this paper was an effective method for synthetic aperture radar target recognition.