摘要
本文首先利用Ｈｏｐｆｉｅｌｄ线性规划神经网络实现了一种极化椭圆匹配技术，从极化分集的雷达后向散射测量结果中提取了目标主散射体的全极化信息──极化分集特征；然后将实时获取的极化分集特征进行特征变换，进而直接利用基于欧氏测度的ＡＲＴ２神经网络进行自动目标识别．
This paper presents a method that maps the polarization-diverse features extraction problem onto the Liapunov energy function of the Hopfield linear programming neural network,so as to obtain a real-time solution of these features derived from the polarization ellipse corresponding to the major scattering centers.Then those features are transformed so that we can make use of the ART2 neural network freely to perform automated target identification.
出处
《电子学报》
EI
CAS
CSCD
北大核心
1994年第3期113-116,119,共5页
Acta Electronica Sinica
关键词
神经网络
特征
目标识别
航空雷达
Linear programming neural network
Polarization-diverse features
Real-time feature extracTion
Automated target identification
