SYNTHETIC APERTURE RADAR IMAGING TECHNOLOGY IN DEEP LEARNING: NEW TRENDS AND VIEWPOINTS
Publication Date: Vol. 32, Issue 2
Submission Deadline: Submissions open from 1 March 2022 through 1 July 2022.
Guest Editors
Achyut Shankar
Amity University
India
ashankar1@amity.edu
Li Zhang
University of London Egham
United Kingdom
Li.Zhang@rhul.ac.uk
Yu Chen Hu
Providence University
Taiwan
ychu@pu.edu.tw
Prabhishek Singh
Amity University
India
psingh29@amity.edu
Scope
The advancement of deep learning has transformed the way to several SAR image processing tasks. The information collected for detecting and tracking ships, ocean wave forecasting, agricultural monitoring, military systems, assessment of damages after flood and earthquake etc. are attained through SAR images. The SAR image quality determines the appropriate information retrieval. The large wavelength and penetrating capability of SAR sensors allow them to acquire images in all weather and during day or night, but the random and continuous interaction of high frequency electromagnetic radiations emitted from SAR sensors with target areas causes constructive and destructive interference, resulting in speckle noise that adversely affects the acquired SAR image. The extraction of information in such a scenario is a difficult task. Apart from speckle noise, SAR images are also affected by geometric distortion, system nonlinear effects, and range migration that needs to be researched as well. There are three different modes of SAR based on nature of their application: strip mapping mode SAR, specifically used for capturing large terrain of area; spotlight mode SAR used for capturing small terrain area by staring at an exact scene from different locations; and inverse SAR used for monitoring the movement of target in war applications. Deep learning methods like a convolutional neural network (CNN) generates incredible results for image classification and restoration purposes. Therefore, new SAR image processing approaches must be created, and SAR raw signal modelling techniques must be developed to help experts, academicians, researchers, and scientists build new SAR systems.
The major objectives of this special section are to:
- Identify the basic research issues related to SAR image processing that are vital for real-world SAR and other remote sensing applications using deep learning technique.
- Monitor the progressive performance made in the solution of remote sensing problems.
- Have experts, academicians, researchers, and scientists share their achievement stories of applying advanced deep learning techniques to the real-world SAR and other remote sensing problems.
We invite manuscripts that successfully apply unconventional and unsupervised deep learning based SAR image processing techniques to various SAR image classification and restoration problems as discussed below. Our topics of interest are broad, including but not limited to the related sub-topics listed below:
- SAR image despeckling using deep learning
- Strip mapping mode SAR image processing using deep learning
- Spotlight mode SAR image processing using deep learning
- Inverse SAR image processing using deep learning
- Solving the problem of geometric distortion in SAR images using deep learning
- Solving the problem of system nonlinear effects in SAR images using deep learning
- Solving the problem of range migration in SAR images using deep learning
- High performance computing for SAR data processing
- Interferometric and polarimetric SAR processing methods
- Electromagnetic scattering models for SAR signal simulation
- Fusion of information from SAR images
- Detection and tracking of ships using deep learning
- Detection of oil natural leakage using deep learning
- Ocean wave forecasting and marine climatology using deep learning
- Regional ice monitoring using deep learning
- Forestry and agricultural land monitoring using deep learning
- Assessment of damage caused by natural calamities such as floods and earthquakes using deep learning
- Detection of small surface movement caused by earthquakes, landslides, or glacier advancement using deep learning
Manuscripts should conform to the author guidelines of the Journal of Electronic Imaging. Prospective authors should submit an electronic copy of their manuscript through the online submission system at https://jei.msubmit.net. The special section should be mentioned in the cover letter. Each manuscript will be reviewed by at least two independent reviewers. Peer review will commence immediately upon manuscript submission, with a goal of making a first decision within six weeks. Each paper is published as soon as the copyedited and typeset proofs are approved by the author.