摘要
为弥补单一模型在识别低空飞行目标时的不足,进一步提高低空飞行目标的识别率,提出一种基于HMM和SVM混合结构的低空飞行目标声识别算法。针对战场环境下声信号的特点,算法综合考虑HMM适合处理连续动态信号及SVM小样本情况下的强分类能力,利用HMM处理待辨识的连续动态信号,将HMM易混淆的信号作为与待辨识信号较为相似的模式类,形成候选模式集,再由SVM在候选模式中对待辨识信号作最后决策。实际数据的识别结果表明相对于单一的HMM和SVM,混合模型的识别率有一定的提高。
In order to overcome the deficiency of the single model for the recognition of low-altitude flying target and improve the recognition rate, a new algorithm is proposed, which takes the advantages of Hidden Markov Model (HMM) and Support Vector Machines (SVM). HMM is good at dealing with sequential inputs, while SVM shows superior performance in classification especially for limited samples. Therefore, they can be combined to get a better and effective multilayer architecture classifier. SVM is used to resolve the uncertainty of the remaining signal which is confusable after the HMM-based recognition. Experimental results prove that the hybrid model has a better performance than the simple one.
出处
《探测与控制学报》
CSCD
北大核心
2009年第6期33-37,共5页
Journal of Detection & Control
基金
2008年度国家自然科学基金项目资助(60872113)