基于支持向量机的空中目标大类别分类

针对已有空中目标识别方法存在的经验风险大、识别率低等不足,依据空中目标的分类原则和纠错码设计原则,设计了针对该问题的纠错码,并训练了码位分类器,最后给出了基于支持向量机的空中目标大类别分类算法。该方法采用纠错编码支持向量机的多类分类技术,降低了经验风险,能对误差进行自动修正,有效地提高了识别率和识别速度。最后给出了一个算例,结果证实了该算法的有效性,并给出了与同类算法的比较结果。

Multi-class classification of air targets based on support vector machine

A new method of multi-class classification of air targets based on support vector machine(SVM) which overcomes defects of existing air targets identification measures effectively is put forward.The SVM(adopts) technique of error-correcting codes(ECC),which can solve the problem easily and fast.Codes classification implement is trained with SVM,and steps of classification method which adopts multi-class classification technical based on ECC-SVM are given simultaneity.The new method can amend error automatically,reduces the experience risk,and improves the rate of identification.At last an example is given,and the computed results are accordant with the experts' advice.