摘要
利用谐振区RCS特征对舰船目标进行识别。深入研究了频率预先优化选择问题,提出了一种新的基于最小分类错误准则的频率选择方法,用以改善目标识别性能。给出了一种基于多类目标假设检验理论的带有拒绝判定目标出现门限的近邻分类器。对五类舰船目标识别的仿真结果表明,新的选频方法显著提高了识别性能;扩展的近邻分类器表现出较为理想的拒判能力。
Ship targets are identified using RCS features of resonance regions. In order to improve the target identification performance, a radar working frequency optimization selection is investigated. A new optimization selection method is given based on the minimum classification error (MCE) rule. A extended nearest neighbor classifier having a rejected threshold is presented based on the muhi type hypothesis testing theory. Target identification simulations indicate that the new frequency optimization method improves the identification performance observably, and the extended nearest neighbor classifier displays good reject ability.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2008年第9期1603-1605,共3页
Systems Engineering and Electronics
基金
航天创新基金资助课题
关键词
船目标识别
频率优选
拒判门限
谐振区
ship target identification
frequency optimization
reject threshold
resonance region