一种基于BP神经网络的高分辨雷达目标识别方法

An Approach to High Resolution Radar Target Recognition Based on BP Neural Network

雷达高分辨距离像(high resolution range profile,HRRP)具有方位敏感性、平移敏感性以及在特征空间高度交叠的特点。利用多维特征,结合神经网络构建的基于神经网络的高分辨雷达目标识别方法,再通过高分辨雷达回波信号的预处理,多种特征值的提取,BP神经网络的合理使用,可有效地提高识别率。

Radar high resolution range profile （HRRP） is sensitive to the target aspect and time-shift variation, and highly overlapped in feature space between different represented by means of multi-dimension features.Based targets, which determines that the target＇s property is well on BP Neural Network and multi-dimension features, an approach to high resolution radar target recognition is proposed in the paper. Preprocessing of echo signal of radar high resolution range, extraction of multi-dimension features and rational use of BP Neural Network,canimprove the target classification performance.The experiment result demonstrates that the method is efficient.