Enriching SAR Ship Detection via Multistage Domain Alignment

The advent of deep learning has made a significant advance in ship detection in synthetic aperture radar (SAR) images. However, it is still challenging since the amount of labeled SAR samples for training is not sufficient. Moreover, SAR images are corrupted by speckle noise, making them complex and...

Full description

Saved in:
Bibliographic Details
Published in:IEEE geoscience and remote sensing letters 2022, Vol.19, p.1-5
Main Authors: Jeong, Somi, Kim, Youngjung, Kim, Sungho, Sohn, Kwanghoon
Format: Article
Language:English
Subjects:
Alignment
Convolutional neural network (CNN)
Datasets
Deep learning
Detection
Domains
Feature extraction
Feature maps
Image resolution
Machine learning
Marine vehicles
Proposals
Radar detection
Radar imaging
Radar polarimetry
Representations
SAR (radar)
ship detection
Speckle
supervised domain adaptation
Synthetic aperture radar
synthetic aperture radar (SAR) image
Task analysis
Training
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The advent of deep learning has made a significant advance in ship detection in synthetic aperture radar (SAR) images. However, it is still challenging since the amount of labeled SAR samples for training is not sufficient. Moreover, SAR images are corrupted by speckle noise, making them complex and difficult to interpret even by human experts. In this letter, we propose a novel SAR ship detection framework that leverages label-rich electro-optical (EO) images for more plentiful feature representations, and delicately addresses the speckle noise in SAR images. To this end, we first introduce a multistage domain alignment module that reduces the distribution discrepancies between EO and SAR feature maps at local, global, and instance levels. This allows enriching SAR representations by gradually instilling cross-domain knowledge from a large-scale EO image dataset. We further design a blind-spot layer for feature extraction to suppress the influence of speckles. Experimental results on the high-resolution SAR images dataset (HRSID) show that our detection performance achieves average precision (AP) 5.5% better than the current state-of-the-arts that exploits SAR images only. Our method significantly improves the detection performance with higher speckle noises, demonstrating stronger robustness than the conventional methods.
ISSN:1545-598X
1558-0571
DOI:10.1109/LGRS.2021.3115498