Loading [a11y]/accessibility-menu.js
CV-CFUNet: Complex-Valued Channel Fusion UNet for Refocusing of Ship Targets in SAR Images | IEEE Journals & Magazine | IEEE Xplore

CV-CFUNet: Complex-Valued Channel Fusion UNet for Refocusing of Ship Targets in SAR Images


Abstract:

In a synthetic aperture radar (SAR) system, target rotation during the coherent integration time results in a time-varying Doppler frequency shift and a blurred image. Bl...Show More

Abstract:

In a synthetic aperture radar (SAR) system, target rotation during the coherent integration time results in a time-varying Doppler frequency shift and a blurred image. Blurred images are not conducive to subsequent information interpretation. This article proposes a complex-valued channel fusion U-shape network (CV-CFUNet) for the 3-D rotation refocusing task of ship targets. The proposed method integrates the refocusing task into a blind inverse problem. To take advantage of the amplitude and phase information of complex SAR images, all elements of CV-CFUNet, including convolutional layer, activation function, feature maps, and network parameters, are extended to the complex domain. The proposed CV-CFUNet is designed by adopting a complex-valued encoder (CV-Encoder), channel fusion module (CFM), and complex-valued decoder (CV-Decoder) to adaptively learn complex features. Experiments on simulated data, GF-3 data, and Sentinel-1 data show that the proposed method achieves significant improvements over existing methods in both efficiency and refocusing accuracy.
Published in: IEEE Transactions on Aerospace and Electronic Systems ( Volume: 59, Issue: 4, August 2023)
Page(s): 4478 - 4492
Date of Publication: 06 June 2023

ISSN Information:

Funding Agency:


References

References is not available for this document.