Unsupervised SAR and Optical Image Matching Using Siamese Domain Adaptation

Zhang, Zhaoxiang; Xu, Yuelei; Qi, Cui; Zhou, Qing; Ma, Lifeng

doi:10.1109/tgrs.2022.3170316

Cited by 13 publications

(5 citation statements)

References 59 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since the work of [5], the deep learning based sparse feature point matching approaches have been extensively explored [20][21][22][23][24][25][26][27][28][29][30][31][32][33][34] for optical and SAR image registration. In most of these researches, the sparse correspondence feature points are identified by extracting the deep feature vector from the local image patch (mostly sized of 100 × 100 to 200 × 200 pixels), and then the best matching is located through local searching.…”

Section: A Deep Learning Based Sparse Feature Point Matching For Opti...mentioning

confidence: 99%

“…Similar approach is applied in [33] which also firstly denoise the SAR images for the subsequent optical-SAR patch matching. In order to deal with the lack of training data problem, the authors of [34] propose to transfer the deep matching models trained with annotated source domains to non-annotated target domains, so as to increase the generalization of the learned models.…”

Section: A Deep Learning Based Sparse Feature Point Matching For Opti...mentioning

confidence: 99%

“…However for optical-SAR image pairs, especially the high resolution ones, a high ratio of mismatches is always obtained, caused by the vast modal disparity in-between. Therefore in recent years, numerous researches have been conducted in the field of optical and SAR sparse feature point matching problem so as to increase the correctly matching rate, including the traditional handcrafted approaches [11][12][13][14][15][16][17][18][19], as well as the learning based methods [20][21][22][23][24][25][26][27][28][29][30][31][32][33][34].…”

Section: Introductionmentioning

confidence: 99%

“…The traditional optical flow algorithms always require brightness constancy assumption, which apparently does not hold for high resolution optical and SAR image pairs. Fortunately, the deep learning technique has shown great potential to learn homogenous features from heterogeneous image pairs [5,[20][21][22][23][24][25][26][27][28][29][30][31][32][33][34]. Furthermore, the deep learning based optical flow methods have recently outperformed the best elaborately designed traditional methods, and also been significantly faster at inference time.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Optical and SAR Image Dense Registration Using a Robust Deep Optical Flow Framework

Zhang

Lei

et al. 2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

The co-registration of optical and SAR imageries is the bottleneck in exploring the complementary information from the two multi-modal datasets. The difficulties lie in not only the complex radiometric relationship between them, but also the distinct geometrical models of the optical and SAR imaging systems, which cause it nontrivial to explicitly depict the spatial relationship between the corresponding image regions when elevation fluctuations exist. This article aims to investigate the optical flow technique for the pixelwise dense registration of the high resolution optical and SAR images, so as to get rid of the outlier removal and geometric mapping procedures, which have to be conducted in the classical image registration approaches that are based on sparse feature point matching. Herein, a deep optical flow framework is designed. Firstly, a dilated feature concatenation method is proposed to enhance the discriminability of the pixelwise features for similarity measurement. An effective network training strategy is used, based on a smoothed flow loss, and also a training dataset that contains simulated elevation fluctuations. Secondly, we propose a self-supervised optical flow finetuning strategy. It incorporates the strength of the blockwise matching approach, which produces better matching precision, into the proposed pixelwise matching procedure. In this way, the accuracy of the optical-SAR dense registration is substantially improved. Extensive experiments conducted on the 1m resolution optical-SAR image pairs of different land-cover types and distinct topographic conditions indicate that, the proposed OSFlowNet-Ft framework is quite robust, and has the potential to perform the optical-SAR image dense registration in practical applications. The Python code of the proposed deep optical flow network will be made available.

show abstract

Section: A Deep Learning Based Sparse Feature Point Matching For Opti...mentioning

confidence: 99%

Section: A Deep Learning Based Sparse Feature Point Matching For Opti...mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Optical and SAR Image Dense Registration Using a Robust Deep Optical Flow Framework

Zhang

Lei

et al. 2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…The Siamese network can also be used to identify corresponding patches in SAR and optical images [23]. Zhang et al [24] provided a method to match unsupervised SAR and optical images by using Siamese domain adaptation. Du et al [25] proposed FM-CycleGAN to achieve feature matching consistency.…”

Section: Introductionmentioning

confidence: 99%

A Fast Matching Method for the SAR Images with Large Viewing Angles Based on Inertial Navigation Information and Neighborhood Structure Consensus

Yan,

Zhao,

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

In the field of multi-view SAR target location, the greater the difference in viewing angles, the higher the target location accuracy. However, this makes it difficult to match the same target between the SAR images with different viewing angles. The performance of traditional SAR image-matching algorithms will deteriorate or even fail to match the images correctly when the viewing angle is gradually increased. To solve this problem, a fast SAR matching method for the SAR images with large viewing angles based on inertial navigation information and neighborhood structure consensus (ININSC) is proposed in this paper. In this algorithm, the key targets are detected in the SAR images by using the maximum connected domain algorithm and the K-means clustering algorithm, and the connected domain centroid of the target is taken as a feature point. The approximate position of the key targets after the viewing angle change is found through inertial navigation information, and then accurate and fast matching is achieved by using the consensus of the neighborhood topological structure of the key targets. The measured data sets demonstrate that compared with traditional SAR image-matching algorithms, the proposed ININSC algorithm solves such a problem that SAR images cannot be accurately matched under the differences at large viewing angles and has better robustness and timeliness.

show abstract

SAR-Optical Image Matching Model Based on Contrastive Learning

Yaoting

2023

2023 20th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)

View full text Add to dashboard Cite

Unsupervised SAR and Optical Image Matching Using Siamese Domain Adaptation

Cited by 13 publications

References 59 publications

Optical and SAR Image Dense Registration Using a Robust Deep Optical Flow Framework

Optical and SAR Image Dense Registration Using a Robust Deep Optical Flow Framework

A Fast Matching Method for the SAR Images with Large Viewing Angles Based on Inertial Navigation Information and Neighborhood Structure Consensus

SAR-Optical Image Matching Model Based on Contrastive Learning

Contact Info

Product

Resources

About