基于调制特征的飞机目标自动分类  

Automatic classification of aircraft based on modulation features

作　　者：丁建江[1] 张贤达[1] 

机构地区：[1]清华大学自动化系,北京100084

出　　处：《清华大学学报（自然科学版）》2003年第7期887-890,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(69772023)

摘　　要：为了解决防空雷达飞机目标自动分类问题。依据飞机旋转部件对雷达电磁波的调制机理、调制特性模型和防空雷达回波特殊性,在时域用相对量分析法提取了幅度、相位调制特征,在频域用复AR双谱分析法提取了调制周期特征。提出了基于幅度、相位。A hi erarchical nearest neighbor classification method was developed based on the thr ee modulation features of amplitude, phase and period to automatically classify aircraft in air-defense radar systems. The relatively constant amplitude and ph ase modulation are extracted using the relative change in the time domain, while the relatively constant periodic modulation is extracted using the complex AR b ispectrum in the frequency-domain based on the induction mechanism, the modulat ion model and the return characteristics. Three types of targets, including eigh t kinds of aircrafts from different manufactures, were classified based on the h ierarchical nearest neighbor classifier with the three modulation features. The classification effectiveness of each modulation feature was analyzed. Experiment al results demonstrate that the three modulation features and the hierarchical n earest neighbor classification method can provide effective automatic classifica tion of aircraft in air-defense radar systems.

关 键 词：防空雷达 飞机目标自动分类 调制幅度 调制相位 调制周期 逐级最近邻分类法 飞机识别 

分 类 号：TN959.17[电子电信—信号与信息处理]