| 神经网络距离-多普勒超分辨雷达成像 |
| |
| 引用本文： | 唐劲松,朱兆达,蒋兴舟.神经网络距离-多普勒超分辨雷达成像[J].海军工程大学学报,1997(3). |
| |
| 作者姓名： | 唐劲松 朱兆达 蒋兴舟 |
| |
| 作者单位： | 海军工程学院海军兵器工程系
(唐劲松)，南京航空航天大学
(朱兆达)，海军工程学院海军兵器工程系(蒋兴舟) |
| |
| 基金项目： | 国家863计划资助项目(308专题) |
| |
| 摘 要： | 研究利用神经网络实现超分辨雷达成像。该方法首先根据Tuff的前后向线性预测(FBLP)方程用神经网络得到样本序列的AR模型参数,然后利用神经网络作线性预测外推,再用神经网络对外推后的数据作离散哈特来变换(DHT)。从哈特来变换可以方便地得到外推数据的离散傅里叶变换,从而得到距离-多普勒超分辨雷达像。利用微波暗室数据对这一方法进行了数字仿真,结果表明这种方法是可行的,即使在低信噪比下也有较好的性能。
|
| 关 键 词： | 雷达成像 神经网络 超分辨 |
A Superresolution Range-Doppler Imaging Method via Neural Network Completely |
| |
| Tang Jinsong Zhu Zhaoda Jiang Xingzhou.A Superresolution Range-Doppler Imaging Method via Neural Network Completely[J].Journal of Naval University of Engineering,1997(3). |
| |
| Authors: | Tang Jinsong Zhu Zhaoda Jiang Xingzhou |
| |
| Institution: | Tang Jinsong Zhu Zhaoda Jiang Xingzhou |
| |
| Abstract: | A method of superresolution range-Doppler imaging via neural network completely is presented in this paper. It is characterized by the following: (1) The FBLP equation is introduced to surperresolution imaging. (2) The Tank-Hopfield network is used to solve FBLP equation. (3) The processes of extrapolating data and DFT are finished by a simple network, which is one part of T-H network. The digital simulation shows that the method is valid. |
| |
| Keywords: | Radar imaging Neural network Superresolution |
| 本文献已被 CNKI 等数据库收录！ |
