Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep Learning Meets SAR

Xiao Xiang Zhu, Sina Montazeri, Mohsin Ali, Yuansheng Hua, Yuanyuan Wang, Lichao Mou, Yilei Shi, Feng Xu, Richard Bamler

Abstract

Deep learning in remote sensing has become an international hype, but it is mostly limited to the evaluation of optical data. Although deep learning has been introduced in Synthetic Aperture Radar (SAR) data processing, despite successful first attempts, its huge potential remains locked. In this paper, we provide an introduction to the most relevant deep learning models and concepts, point out possible pitfalls by analyzing special characteristics of SAR data, review the state-of-the-art of deep learning applied to SAR in depth, summarize available benchmarks, and recommend some important future research directions. With this effort, we hope to stimulate more research in this interesting yet under-exploited research field and to pave the way for use of deep learning in big SAR data processing workflows.

Deep Learning Meets SAR

Abstract

Deep learning in remote sensing has become an international hype, but it is mostly limited to the evaluation of optical data. Although deep learning has been introduced in Synthetic Aperture Radar (SAR) data processing, despite successful first attempts, its huge potential remains locked. In this paper, we provide an introduction to the most relevant deep learning models and concepts, point out possible pitfalls by analyzing special characteristics of SAR data, review the state-of-the-art of deep learning applied to SAR in depth, summarize available benchmarks, and recommend some important future research directions. With this effort, we hope to stimulate more research in this interesting yet under-exploited research field and to pave the way for use of deep learning in big SAR data processing workflows.

Paper Structure

This paper contains 28 sections, 10 figures, 1 table.

Table of Contents

  1. Motivation
  2. Introduction to Relevant Deep Learning Models and concepts
  3. Deep Learning Models
  4. Convolutional Neural Networks (CNN)
  5. Recurrent Neural Networks (RNN)
  6. GANs
  7. Supervised, Unsupervised and Reinforcement Learning
  8. Supervised Learning
  9. Unsupervised Learning
  10. Deep Reinforcement Learning (DeepRL)
  11. Relevant Deep Learning Concepts
  12. Automatic Machine Learning (AutoML)
  13. Geometric Deep Learning -- Graph Neural Networks (GNNs)
  14. Possible Pitfalls
  15. Recent Advances in Deep learning applied to SAR
  16. ...and 13 more sections

Figures (10)

  • Figure 1: A Selection of relevant deep learning models. Sources of the images: VGG ferguson_automatic_2017, ResNet online_resnet, U-Net online_unet, LSTM online_lstm, RNN feng_audio_2017, VAE online_vae, GAN online_gan, CGNN Zitnik2018, RGNN huang_residual_2019, and DeepRL zoph_neural_2017.
  • Figure 2: Classification maps obtained from a TerraSAR-X image of a small area in Norway geng2017deep. Subfigures (a)-(f) depict the results of classification using SVM (accuracy = 78.42%), sparse representation classifier (SRC) (accuracy = 85.61%), random forest (accuracy = 82.20%) uhlmann2014integrating, SAE (accuracy = 87.26%) xie2014multilayer, DCAE (accuracy = 94.57%) geng2015high, contractive AE (accuracy = 88.74). Subfigures (g)-(i) show the combination of DSCNN with SVM (accuracy = 96.98%), with SRC (accuracy = 92.51%) hou_sar_2016, and with random forest (accuracy = 96.87%). Subfigures (j) and (k) represent the classification results of DSCNN (accuracy = 97.09%) and DSCNN followed by spatial regularization (accuracy = 97.53%), which achieve higher accuracy than the other methods.
  • Figure 3: The architecture of the dual-branch deep convolution neural network (Dual-CNN) for PolSAR image classification, proposed in gao2017dual.
  • Figure 4: The flowchart of the multi-aspect-aware bi-directional approach for SAR ATR proposed in zhang2017sar.
  • Figure 5: Very high resolution TerraSAR-X image of Berlin (left), and the predicted building mask shahzad2019buildings (right).
  • ...and 5 more figures