
端到端全复数域SAR目标分类神经网络
End-to-End Full Complex-Valued Domain SAR Target Classification Neural Network
合成孔径雷达(Synthetic Aperture Radar,SAR)图像探测经常遇到误差敏感和计算量大的问题,给SAR图像目标识别带来困难,为此研究人员针对SAR数据提出很多新颖高效的深度学习方法,但面向SAR目标识别的深度学习网络大多使用与光学实值处理相同的方法,直接将实值深度神经网络应用于SAR图像.实值的神经网络都不同程度丢失了相位信息,不能充分利用SAR数据复数特性.相位信息是SAR图像独有的数据特征,在SAR干涉测量、信息检索、目标识别等应用中起到至关重要作用.本文为了使网络更适合SAR复数数据特征提取,打破传统神经网络架构,将整体网络端到端复数化,提出一种新型的端到端全复数域多层级神经网络(Complex-valued mUltI-Stage convolutIonal Neural nEtworks,CUISINE)架构,从SAR复数图像数据的输入到卷积计算,再到分类标签,实现全网络复数域下的计算.通过在MSTAR(Moving and Stationary Target Acquisition and Recognition)公开数据集上实验对比表明,本文的方法在SAR数据目标分类上有很好表现,在相位误差为0 rad的测试集上正确率达到99.42%,在相位误差为50 rad的测试集上正确率达到88.05%.
Synthetic aperture radar (SAR) image detection often encounters problems such as error sensitivity and high computational complexity, which pose challenges to SAR target recognition. Researchers have proposed many novel and efficient deep learning methods for SAR data. However, most of these deep learning networks for SAR target recognition use the same methods as optical real-valued processing, directly applying real-valued deep neural networks to SAR images. Real-valued neural networks to some extent lose the phase information, which cannot fully utilize the complex characteristics of SAR data. As phase information is a unique data feature in SAR images, it plays a crucial role in applications such as SAR interferometry, information retrieval, and target recognition. In order to make the network more suitable for extracting complex data features from SAR, breaking the architecture of traditional neural networks, this paper proposes a novel end-to-end fully complex-valued multi-stage convolutional neural network (Complex-valued mUltI-Stage convolutIonal Neural nEtworks, CUISINE) architecture. It realizes the computation in the full complex-valued domain from the input of SAR complex image data to convolutional calculations, and finally to classification labels. Experimental comparisons on the publicly available MSTAR dataset show that our method performs well in SAR target classification. The accuracy reaches 99.42% on the test set with a phase error of 0 rad, and 88.05% on the test set with a phase error of 50 rad.
合成孔径雷达 / 相位信息 / 全复数域 / 端到端 / 神经网络 {{custom_keyword}} /
synthetic aperture radar / phase information / full complex fields / end-to-end / neural networks {{custom_keyword}} /
表1 10类目标测试集与训练集 |
类别 | 训练集 | 测试集 | ||
---|---|---|---|---|
样本数量 | 俯仰角/(°) | 样本数量 | 俯仰角/(°) | |
2S1 | 299 | 17 | 274 | 15 |
BMP2 | 233 | 17 | 196 | 15 |
BDRM2 | 298 | 17 | 274 | 15 |
BTR60 | 256 | 17 | 195 | 15 |
BTR70 | 233 | 17 | 196 | 15 |
D7 | 299 | 17 | 274 | 15 |
T62 | 299 | 17 | 273 | 15 |
T72 | 232 | 17 | 196 | 15 |
ZIL131 | 299 | 17 | 274 | 15 |
ZSU23/4 | 299 | 17 | 274 | 15 |
表2 训练集与测试集数据规模 |
相位误差/rad | 0 | 10 | 15 | 20 | 25 | 30 | 35 | 40 | 50 |
---|---|---|---|---|---|---|---|---|---|
训练集 | 2 747 | 2 747 | 2 747 | 2 747 | 2 747 | — | — | — | — |
总计 | 13 735 | ||||||||
测试集 | 2 426 | 2 426 | 2 426 | 2 426 | 2 426 | 2 426 | 2 426 | 2 426 | 2 426 |
表3 有相位误差训练集对比实验 (%) |
相位误差/rad | 0 | 10 | 15 | 20 | 25 | 30 | 35 | 40 | 50 |
---|---|---|---|---|---|---|---|---|---|
CNN1 | 97.24 | 97.07 | 96.87 | 95.71 | 93.24 | 88.09 | 79.31 | 69.33 | 54.70 |
A-ConvNet1 | 97.20 | 96.13 | 95.57 | 94.11 | 90.60 | 79.97 | 65.50 | 47.40 | 23.30 |
FEN1 | 99.63 | 99.09 | 99.05 | 98.56 | 93.95 | 94.77 | 87.68 | 76.01 | 50.04 |
CCNN2 | 98.19 | 97.61 | 97.94 | 97.57 | 96.13 | 93.03 | 87.55 | 79.43 | 62.94 |
MS-CVNets2 | 98.02 | 98.31 | 98.39 | 98.52 | 97.28 | 95.42 | 91.14 | 85.45 | 69.47 |
CV-Net2 | 98.85 | 98.97 | 99.13 | 99.13 | 98.47 | 97.11 | 94.31 | 89.16 | 74.69 |
CUISINE* | 99.42 | 98.93 | 98.89 | 99.05 | 98.35 | 97.69 | 96.58 | 94.27 | 88.05 |
表4 消融实验 (%) |
模型结构/rad | 0 | 10 | 15 | 20 | 25 | 30 | 35 | 40 | 50 |
---|---|---|---|---|---|---|---|---|---|
复数卷积 | 92.87 | 93.73 | 93.69 | 92.13 | 89.74 | 85.28 | 79.72 | 74.53 | 62.08 |
复数卷积+可分离 | 96.70 | 96.78 | 95.96 | 95.26 | 94.19 | 92.00 | 89.86 | 86.44 | 78.19 |
复数卷积+可分离+多尺度 | 97.90 | 97.69 | 97.16 | 97.03 | 96.00 | 94.23 | 92.70 | 89.78 | 80.87 |
复数卷积+可分离+多尺度+复数标签* | 99.42 | 98.93 | 98.89 | 99.05 | 98.35 | 97.69 | 96.58 | 94.27 | 88.05 |
1 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
2 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
3 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
4 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
5 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
6 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
7 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
8 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
9 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
10 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
11 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
12 |
张智, 易华挥, 郑锦. 聚焦小目标的航拍图像目标检测算法[J]. 电子学报, 2023, 51(4): 944-955.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
13 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
14 |
郭晓轩, 冯其波, 冀振燕, 等. 多线激光光条图像缺陷分割模型研究[J]. 电子学报, 2023, 51(1): 172-179.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
15 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
16 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
17 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
18 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
19 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
20 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
21 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
22 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
23 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
24 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
25 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
26 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
27 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
28 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_ref.label}} |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
/
〈 |
|
〉 |