当前位置: X-MOL 学术Radio Sci. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Optimizing Forward‐Looking Missile‐Borne BSAR Imaging With Convolutional Neural Network
Radio Science ( IF 1.6 ) Pub Date : 2021-02-03 , DOI: 10.1029/2020rs007248
Hongyin Shi 1, 2 , Yupeng Yang 1, 2 , Chunyang Lv 1, 2 , Jing Zhang 1, 2 , Jianwen Guo 1, 2
Affiliation  

During the guidance of the missile, the target needs to be positioned in real time. The detonating missile will look forward to the target area during the flight, and the imaging distance becomes closer as the missile approaches the target. The image quality of forward‐looking missile‐borne bistatic synthetic aperture radar (FM‐BSAR) deteriorates rapidly when the detonator enters a rapid dive stage. One of the most serious problem is the azimuth defocusing, the more deviated from the center of the imaging aperture, the more defocusing. In this paper, a method of using convolutional neural network (CNN) to accomplish image focusing is proposed to improve the image quality of BSAR. The advantage of this method is to focus the image accurately with inaccurate imaging parameters. The experimental results show that the proposed method can accurately locate and focus the defocusing points, which has practical significance for improving the BSAR close‐range imaging effect of high‐maneuvering platform.

中文翻译:

卷积神经网络优化前瞻性导弹波音BSAR成像

在导弹制导过程中,需要实时定位目标。爆炸中的导弹将在飞行过程中期待目标区域,并且随着导弹接近目标,成像距离变得更近。当雷管进入快速潜水阶段时,前瞻性导弹双基地合成孔径雷达(FM-BSAR)的图像质量会迅速下降。最严重的问题之一是方位角散焦,离成像孔径中心越远,散焦就越多。提出了一种利用卷积神经网络(CNN)进行图像聚焦的方法,以提高BSAR的图像质量。此方法的优点是使用不准确的成像参数精确聚焦图像。
更新日期:2021-02-26
down
wechat
bug