高分辨一维距离像(High Resolution Range Profile,HRRP)包含了丰富的目标信息,通过提取HRRP的强散射点分布特征可实现对不同目标的分类和识别.论文充分考虑了HRRP的多域特征和时间依赖性,利用ResNet18网络进行时频域特征提取,并结合记忆融合网络(Memory Fusion Network,MFN)提出新型深度学习模型MI-MFN(Multi Input-MFN)进行多域特征的融合识别,实现了在不同维度的记忆的跨视图交互,有效地学习和提取HRRP序列特征.实验结果表明,MI-MFN模型的识别准确率可以达到99.9%以上,具有出色的识别性能.
HRRP Target Recognition Method Based on Multi Domain Feature Fusion
The High Resolution Range Profile contains rich target information,and by extracting the strong scattering point distribution features of HRRP,different targets can be classified and recognized.This article fully considers the multi domain fea-tures and time dependence of HRRP,utilizes ResNet18 network for time-frequency domain feature extraction,and proposes a new deep learning model MI-MFN combined with memory fusion network for multi domain feature fusion recognition.It achieves cross view interaction of memory in different dimensions,effectively learning and extracting HRRP sequence features.The experimental results show that the recognition accuracy of the MI-MFN model can reach over 99.9%,and it has excellent recognition perfor-mance.
万方数据
维普
