用于合成孔径雷达中射频干扰检测与分割的轻量级深度神经网络。

Lightweight deep neural network for radio frequency interference detection and segmentation in synthetic aperture radar.

作者信息

Zheng Fenghao, Zhang Zhongmin, Zhang Dang

机构信息

College of Information and Communication Engineering, Harbin Engineering University, Harbin, 150001, China.

出版信息

Sci Rep. 2024 Sep 5;14(1):20685. doi: 10.1038/s41598-024-71775-8.

DOI:10.1038/s41598-024-71775-8

PMID:39237592

原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC11377538/

Abstract

Radio frequency interference (RFI) poses challenges in the analysis of synthetic aperture radar (SAR) images. Existing RFI suppression systems rely on prior knowledge of the presence of RFI. This paper proposes a lightweight neural network-based algorithm for detecting and segmenting RFI (LDNet) in the time-frequency domain. The network accurately delineates RFI pixel regions in time-frequency spectrograms. To mitigate the impact on the operational speed of the entire RFI suppression system, lightweight modules and pruning operations are introduced. Compared to threshold-based RFI detection algorithms, deep learning-based segmentation networks, and AC-UNet specifically designed for RFI detection, LDNet achieves improvements in mean intersection over union (MIoU) by 24.56%, 13.29%, and 7.54%, respectively.Furthermore, LDNet reduces model size by 99.03% and inference latency by 24.53% compared to AC-UNet.

摘要

射频干扰（RFI）给合成孔径雷达（SAR）图像分析带来了挑战。现有的RFI抑制系统依赖于RFI存在的先验知识。本文提出了一种基于轻量级神经网络的算法，用于在时频域中检测和分割RFI（LDNet）。该网络能在时频频谱图中准确勾勒出RFI像素区域。为减轻对整个RFI抑制系统运行速度的影响，引入了轻量级模块和剪枝操作。与基于阈值的RFI检测算法、基于深度学习的分割网络以及专门为RFI检测设计的AC-UNet相比，LDNet的平均交并比（MIoU）分别提高了24.56%、13.29%和7.54%。此外，与AC-UNet相比，LDNet的模型大小减少了99.03%，推理延迟降低了24.53%。