HRRP feature extraction and recognition method of radar ground target using convolutional neural network

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

High resolution range profile(HRRP) is an important feature of radar target. It can be used to classify and recognize target. As for the attitude sensibility, the HRRP of target will change dramatically within a small angle. Since the convolutional neural network(CNN) is good tool to abstract the HRRP features from the original input data, and these features have good generalization ability to represent the target HRRP. Aiming at this issue, an approach is proposed to realize radar ground target HRRP features extraction and recognition based on convolutional neural network. The target scattering center features are reorganized into a 2-D feature map. Then a CNN is constructed to recognize the target based on the feature map of the target. The computation result based on MSTAR database shows the performance of the proposed approach.

Original languageEnglish
Title of host publicationProceedings of the 2019 21st International Conference on Electromagnetics in Advanced Applications, ICEAA 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages658-661
Number of pages4
ISBN (Electronic)9781728105635
DOIs
StatePublished - Sep 2019
Event21st International Conference on Electromagnetics in Advanced Applications, ICEAA 2019 - Granada, Spain
Duration: 9 Sep 201913 Sep 2019

Publication series

NameProceedings of the 2019 21st International Conference on Electromagnetics in Advanced Applications, ICEAA 2019

Conference

Conference21st International Conference on Electromagnetics in Advanced Applications, ICEAA 2019
Country/TerritorySpain
CityGranada
Period9/09/1913/09/19

Keywords

  • CNN
  • Feature extraction
  • HRRP
  • MSTAR

Fingerprint

Dive into the research topics of 'HRRP feature extraction and recognition method of radar ground target using convolutional neural network'. Together they form a unique fingerprint.

Cite this