Phase unwrapping (PU) is a critical step in interferometric synthetic aperture radar (InSAR) data processing. Among all PU methods, multibaseline PU (MBPU) methods is a state-of-the-art method, which can overcome the limitations of the phase continuity assumption in the traditional single-baseline PU methods. However, the MBPU methods still cannot effectively balance PU accuracy and efficiency when processing large-size interferograms. In particular, the current best MBPU methods, two-stage programming approach (TSPA), cannot work well when the baseline ratio is less than 2, limiting its application. To solve this problem, a WaveCluster-based robust and fast MBPU algorithm (WCRFPU) is proposed in this paper. First, appropriate initial grid and neighborhood parameters are selected according to the exclusive information of the InSAR data set to reduce the number of wrong clusters caused by dimension mismatch. Then the WaveCluster algorithm is used to cluster the intercept map with 3D clustering features, which can obtain more accurate clustering results and efficiently handle large-size interferograms. Subsequently, a cluster correction step is added to improve the PU accuracy further. Theoretical analysis and experimental results show that this method has more advantages than the existing MBPU methods in efficiency, accuracy, and adaptability to baseline ratios when processing large-size interferograms.

Phase unwrapping (PU) is a critical process for numerous synthetic aperture radar interferometry (InSAR) applications. The advent of deep learning (DL) has revolutionized PU, with various DL-based methods emerging over the past four years that consistently achieve state-of-the-art results. However, the generalization capability of these methods is impeded by the lack of a large-volume training dataset for InSAR PU. In contrast, model-based PU methods, both single-baseline (SB) and multibaseline (MB), utilize mathematical formulations that represent the PU knowledge, often resulting in robust generalizability in their respective domains. This article introduces MoDL-PU, a hybrid model-based/DL-based PU method. MoDL-PU, leveraging the Res-UNet-Inception architecture, employs a two-stage training approach: initial dataset-based training to extract low-frequency unwrapped phase information, followed by retraining with three specialized PU knowledge to enhance high-frequency details for targeted PU tasks: SB PU (MoDL-SBPU), MB PU for Digital Elevation Model (DEM) reconstruction (MoDL-MBPUT), and MB PU for deformation monitoring (MoDL-MBPUD). The proposed method exhibits robust generalizability, interpretability, and noise resilience, effectively combining the advantages of model-based and DL-based PU methodologies. The experimental outcomes demonstrate that MoDL-PU not only exceeds the performance of conventional model-based PU methods but also competes favorably with state-of-the-art supervised and self-supervised DL-based approaches.

The Hongtu-1 (HT-1) Synthetic Aperture Radar (SAR) constellation is the first in-orbit spaceborne single-pass multi-baseline interferometric SAR (InSAR) system. The system has the ability to conduct high-resolution earth observation and high-precision, high-efficiency terrain surveying. The highest resolution of the system is better than 0.5 m, and it has a 1:50000 scale global digital elevation model (DEM) and digital surface model (DSM) surveying capability. This paper provides a basic introduction to the HT-1 constellation, and demonstrates the advantages of single-pass multi-baseline InSAR results and its advantages over steep area.

Deep learning-based methods have dominated in object detection in synthetic aperture radar (SAR) images. Despite significant advancements, existing approaches mainly focus on architectural enhancements of the network, leaving the unique challenges posed by the strong speckle noise inherent in SAR imagery not fully tackled. In this article, we introduce the multilevel denoising detection transformer (MD-DETR) to mitigate the impact of speckle noise for SAR object detection. We build MD-DETR in three steps, leveraging the advanced real-time detection transformer (RT-DETR) framework. First, we address the noise data input by designing a straightforward image-level denoising technique, which concatenates the original image with its denoised counterparts to create a new image for model input. Subsequently, for feature-level denoising, a coarse-mask guidance feature learning module is proposed to enhance the global features of the decoder. In addition, for query-level denoising, we introduce an auxiliary head with the one-to-many matching strategy. This enables flexible query selection by dynamically adjusting the number of queries to suit various scenarios. To validate the superiority of the proposed MD-DETR, extensive experiments are conducted on two benchmark datasets, that is, the SAR ship detection dataset (SSDD) and the recently published COCO-level large-scale multiclass SAR object detection dataset (SARDet-100K). Experimental results on both datasets outperform previous advanced detectors, achieving a new state-of-the-art (SOTA) with 98.9 AP50 and 88.5 mAP50 on SSDD and SARDet-100K, respectively.

Multibaseline (MB) interferometric synthetic aperture radar (InSAR) is an advanced variant of conventional InSAR that aims to enhance the accuracy and reliability of phase unwrapping (PU). Among the PU methods employed in MB-InSAR, the maximum likelihood (ML) method offers an optimal solution for phase estimation. However, its limited noise robustness has hindered its practical applicability. To address this limitation, we propose a novel approach, named InSAR phase probability density function (PDF)-to-height/deformation (PDF2HD), which leverages a newly introduced deep convolutional neural network (DCNN) with exceptional anti-noise capabilities. The PDF2HD method employs U-Net and residual network to estimate the InSAR PDF, enabling it to mitigate the influence of phase noise. We present experimental results using two simulated MB InSAR datasets to demonstrate the effectiveness of our proposed method for both digital elevation model (DEM) reconstruction and deformation monitoring.

Multibaseline (MB) interferometric synthetic aperture radar (InSAR) is an extension of conventional InSAR and is used to improve phase unwrapping (PU) accuracy without obeying the Itoh condition. The Chinese remainder theorem (CRT) is the mathematical foundation of most MB PU algorithms for determining the connection between different MB interferograms. However, CRT only exploits the relationship between the interferometric phase and terrain height or surface deformation alone, i.e., the phases of topography and deformation are considered as the measurement biases for each other in the traditional processing chains. In other words, traditional MB InSAR cannot directly obtain InSAR products from interferograms that contain both the topography and deformation phases without external information or assumptions. To solve this issue, differing from CRT, this article presents a new mathematical framework for MB PU, which provides a likelihood function to jointly estimate the topography and deformation velocity gradient. Based on this framework, a novel MB PU method is proposed for simultaneously obtaining a digital elevation model (DEM) and deformation information from interferograms. The experimental results show that the proposed method is effective and efficient for DEM reconstruction and deformation monitoring.

A deep convolutional neural network (DCNN) has been widely applied to the 2-D phase unwrapping (PU) in synthetic aperture radar interferometry (InSAR). Our previously-developed PGNet and BCNet outperform the model-based 2-D PU methods. However, the two networks can be further improved. As the PGNet is limited to estimating the phase gradients within ±2π, unwrapped phases can be incorrectly unwrapped sometimes. The BCNet is sensitive to the high-density distribution of the residues caused by a noisy interferogram, resulting in many isolated regions. To solve both issues, we bridge the PGNet and BCNet, studying a new DCNN-based 2-D PU framework (PG-BCNet). The results show that the PG-BCNet is more noise-robust than that of the BCNet and overcomes the limitation of the PGNet that cannot unwrap the phase gradients beyond ±2π.

Deep convolutional neural networks (DCNN)-based methods have been applied widely to ship detection in SAR images. However, most DCNN-based ship target detectors that focus on the detection performance ignore the computation complexity. We propose a lightweight anchor-free ship detection network (LASDNet) for SAR images to tackle this problem. First, a lightweight backbone utilizing a double fusion with squeeze-and-excitation-bottleneck block under the CSPNet design (CSP-DFSEB) and three pooling blocks (i.e., EVE, FCT, and ME blocks) are constructed, which achieves a balance between accuracy and efficiency. Second, a transformer-based aggregation layer conducts feature fusion. Finally, an improved one-stage anchor-free detector FCOS is presented. The analyses of the High-Resolution SAR Images Dataset for Ship Detection and Instance Segmentation (HRSID) dataset show that the proposed detector has the second least number of parameters (1.15 MB), the lowest computation complexity (1.01 GFLOPs), and the highest average precision (59.25) compared with other state-of-the-art methods.

Interferometric phase denoising is a vital procedure for interferometric synthetic aperture radar (InSAR)-based remote sensing techniques because it can improve the accuracy of the final InSAR product. Here, we propose a deep convolutional neural network (DCNN)-based InSAR phase denoising method, abbreviated PDNet. Given an ideal wrapped phase, φ, the PDNet learns the self-similarity function of φ from the input interferogram. After training, the PDNet obtains filtered wrapped phases using the maximum-likelihood approach by exhausting all φs from –π to π. Unlike a boxcar-based filtering method, the PDNet does not consist of an “averaging operation” on the spatial domain, and the resolution loss and interferometric fringe distortion will not directly affect the PDNet result. Thus, the PDNet can be considered a nonlocal phase denoising approach. Analyses and results show that the PDNet is an almost near-real-time denoising algorithm. Its denoising accuracy is higher than that of the available model- and learning-based InSAR phase denoising methods.

Multibaseline (MB) phase unwrapping (PU) is a vital processing procedure for MB synthetic aperture radar interferometry (InSAR) signal processing and can improve the traditional InSAR by changing the ill-posed problem to the well-posed problem. The existing research has shown that the MB PU problem can be successfully converted into an unsupervised cluster analysis problem. Using the high feature descriptiveness of the deep learning technique, an unsupervised deep convolutional neural network, referred to as CANet, is proposed to cluster all the pixels into different groups according to the input’s recognizable pattern of the ambiguity number of the MB interferometric phase. Subsequently, we extend our previous two-stage programming-based MB processing approach (TSPA) to processing MB PU on a sparse irregular network, which is established from the clustering result of CANet. Both theoretical analysis and experimental results show that the proposed method is an effective MB PU method, and its execution time is drastically lower than those of many classical MB PU methods.

Two-dimensional (2-D) phase unwrapping (PU) is a critical processing step for many synthetic aperture radar (SAR) interferometry (InSAR) applications. As is well known, the traditional 2-D PU is an ill-posed inverse problem, which means that regardless of how skillful the PU algorithm designer is, it is impossible to design an algorithm that can correctly process all the 2-D PU situations, i.e., we can only design the best PU algorithm in the statistical sense. Therefore, accumulating PU processing experience from different study cases is important for PU algorithm design. Currently, the deep learning (DL) technique provides a potential framework to accumulate processing experience, and a flood of valuable data coming from different InSAR sensors provides the ability to enable the learning-based PU technique outside the traditional model-based technique. In this article, we transform the 2-D PU problem into a learnable image semantic segmentation problem and propose a DL-based branch-cut deployment method (abbreviated as BCNet). To start, we propose the optimal branch-cut connection criterion (referred to as OPT-BC) with the reference unwrapped phase given. Next, using the relationship between the residue and branch-cut as the learning objective, BCNet is trained using the samples provided by OPT-BC to produce the branch-cut result. Finally, the traditional branch-cut method is utilized to perform the postprocessing procedure to obtain the final PU result. The experimental results demonstrate that the proposed BCNet-based PU method is a near-real-time 2-D PU algorithm, and its accuracy outperforms the traditional model- and learning-based 2-D PU methods.

Two-dimensional phase unwrapping (PU) is a classical ill-posed problem in synthetic aperture radar interferometry (InSAR). The traditional algorithmic model-based 2-D PU methods are limited by the Itoh condition, which is from the PU researchers’ experience and has critical challenges under strong phase noises or violent phase changes. Recently, advanced learning-based 2-D PU methods could break through the limitation of the Itoh condition owing to their data-driven frameworks, offering promising results in terms of both the speed and accuracy. The one-step learning-based PU method, as one of the representatives, retrieves the unwrapped phase directly from the wrapped phase through regression. However, the main disadvantage of one-step learning-based PU is that it usually blurs the output unwrapped phase due to its$L_{2}$loss, that is, it cannot guarantee the congruency between the rewrapped interferometric fringes of the PU solution and the input interferogram. To solve this problem, we propose a one-step 2-D PU method based on the conditional generative adversarial network (referred to as PU-GAN), which treats 2-D PU as an image-to-image translation problem. The generator in PU-GAN can be trained to generate the unwrapped phase through minimizing a$L_{1}$-norm loss based on a U-Net architecture, while simultaneously the corresponding discriminator can learn an adversarial loss by a structure of Patch-GAN that tries to classify if the output unwrapped phase image is real or fake. Both a theoretical analysis and the experimental results show that the proposed method outperforms the representative algorithmic model-based and learning-based 2-D PU methods.

In this paper the effectiveness of a CNN based interferometric phase unwrapping algorithm combined with phase noise filtering is analysed. In particular, the considered processing chain relies on a pre-processing step with the nonlocal filter InSAR-BM3D followed by a deep CNN solution for restoring the absolute phase. The analyses is conducted on simulated data with different coherence values and aims at comparing the performance of the unwrapping with and without the pre-processing step. This paper is the first step towards a unique deep learning solution for jointly unwrapping and restoring the absolute phase.

Synthetic aperture radar (SAR) images are widely used for aerial and spatial image applications. However, Most of SAR images are usually grayscale images with no color information. Hence, the study of SAR image colorization is meaningful. At present, deep learning has become the mainstream method of SAR coloring, and with the most advanced pix2pix method, it achieves satisfactory results. However, such an approach is limited to the corresponding paired data, which may be difficult to get. We then notice that the cycle-consistency loss can remove this constraint to some extent. In this letter, we present a novel method to colorize the SAR image using a multidomain cycle-consistency generative adversarial network (MC-GAN). The proposed method improves the performance of coloring SAR images from two aspects: first, we propose a mask vector for images of every particular terrain combined with cycle-consistency loss, which does not need the paired SAR-optical images to train the model. Second, we define the multidomain classification loss, which can together get the correct output image with the color we hope it to be. We examined the proposed method on the newly SEN1-2 data set compared with the pix2pix and CycleGAN methods, which demonstrates the effectiveness of our proposed method.

Multi-baseline (MB) phase unwrapping (PU) is an important processing stage of MB synthetic aperture radar (SAR) interferometry (InSAR). Compared with the traditional single-baseline (SB) PU, MB PU is more applicable to the mountainous area, which does not follow the Itoh condition. A two-stage programming MB PU approach (TSPA) proposed by H. Yu, which builds the link between SB and MB PUs, so many existing classical SB PU methods can be transplanted into MB domain. In this paper, an extended Goldstein's Branch-cut algorithm for MB InSAR using the TSPA, abbreviated as TSPA-BC, is proposed, consisting of three steps. In step 1, the MB residues are identified according to the phase gradients estimated by TSPA based on CRT. In step 2, the MB branch cuts with global minimal length are generated, which is equal to the MB L° -norm branch-cut length. In step 3, the final PU result are obtained by a flood-fill integration process using the phase gradients obtained by stage 1 in which the integration path does not pass through any MB branch cut. The experimental results of TanDEM-X MB dataset illustrate that the effectiveness of the TSPA-BC method in the rugged area.

Two-dimensional phase unwrapping (2-D PU) is one of the key processes in reconstructing the topography or displacement of the Earth surface from its interferometric synthetic aperture radar (InSAR) data. Estimating the absolute phase gradient information is an unavoidable step utilized by almost all the 2-D PU methods. Traditionally, the gradient estimation step relies on the phase continuity assumption, which requests that the observed area has spatial continuity. However, the abrupt topographic changes and system noise usually results in the failure of the phase continuity assumption in reality. Under this condition, it is difficult for the traditional 2-D PU to provide the correct absolute phase over the area with abrupt interferometric fringe change or with strong system noise. To solve the issue, we propose a novel deep convolutional neural network (DCNN), abbreviated as PGNet, to estimate the phase gradient information instead of the phase continuity assumption in this article. The major advantage of PGNet lies in its deep architecture to learn the characteristics of phase gradients from enormous training images with different noise levels and topographic features. Subsequently, the L1-norm objective function is used to minimize the difference between unwrapped phase gradients and the gradients estimated by PGNet for obtaining the final PU result. Taking the phase gradient pattern of the TerraSAR-X-TanDEM-X interferogram as the learning object, experimental results demonstrate the absolute phase gradient estimated by PGNet is more credible than that from the phase continuity assumption such that the corresponding PU result outperforms those obtained by the traditional 2-D PU methods.

Phase unwrapping (PU) is one of the key process in reconstructing the digital elevation model (DEM) of a scene from its interferometric synthetic aperture radar (InSAR) data. Compared with traditional single-baseline PU, the multibaseline PU does not need to obey the phase continuity assumption, which can be applicable to reconstruct the DEM where topography varies drastically. However, the performance of the multibaseline PU is directly concerned with noise level. Contrarily, the single-baseline PU algorithm has good noise robustness, since it is based on the globe wrapped phase information, such as PU-max-flow (PUMA) algorithm. In order to improve the noise robustness of the multibaseline, in this paper, we extend single-baseline PUMA algorithm to multibaseline domain, referred to as multibaseline PUMA algorithm, which allows the unwrapping of multibaseline interferograms for the generation of DEM. The proposed algorithm does not need to obey the phase continuity assumption by taking the advantages of multibaseline diversity and improves the noise robustness by using the global wrapped information both from single- and multibaseline domain. The performance of the proposed algorithm is tested on simulated InSAR data experiments, which demonstrate the effectiveness and noise robustness of the proposed algorithm.