航空学报 > 1994, Vol. 15 Issue (8): 1007-1011

用TH神经网络方法外推数据的超分辨雷达成象

叶蓁如, 殷军   

  1. 南京航空航天大学电子工程系,南京,210016
  • 收稿日期:1992-08-24 修回日期:1993-11-03 出版日期:1994-08-25 发布日期:1994-08-25

SUPERRESOLUTION RADAR IMAGING WITH EXTRAPOLATING DATA USING NEURAL NETWORK

Yezhenru, YinJun   

  1. Department of Electronic Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing,210016
  • Received:1992-08-24 Revised:1993-11-03 Online:1994-08-25 Published:1994-08-25

摘要: 研究用Tank-Hopfield神经网络(THNN)求解AR模型参数作数据外推的超分辨雷达成象,并用微波暗室实测数据对THNN方法和Burg方法作了验证,结果表明,两种方法均能在较低的信噪比条件下实现超分辨成象,且随着VLSI技术的发展,神经网络方法将是一种很有希望的超分辨成象方法。

关键词: 雷达成象, 高分辨率, 线性预测, 人工智能

Abstract: ata extraploation super-resolution radar imaging with AR parameters estimated by Tank-Hopfield neural network(THNN)is investigated,and the real data in the microwave anechoic chamber are proeessed by the TH neural network method and Burg method.Imaging results indicate that both methods can complete superresolution imaging under the lower SIN,and with the developing of the VLSI technology,the neural network method will be a very promising super resolution imaging method.

Key words: radar imagery, high resolution, linear prediction, artificial intelligence

中图分类号: