High Resolution SAR Images Research Articles

Research on High Resolution SAR Imaging Algorithm Based on Optimal Sampling Length

Research on High Resolution SAR Imaging Algorithm Based on Optimal Sampling Length

Refocusing High-Resolution SAR Images of Complex Moving Vessels Using Co-Evolutionary Particle Swarm Optimization

To increase the global convergence and processing efficiency of particle swarm optimization (PSO) applied in the adaptive joint time-frequency, in this study an improved PSO is proposed to refocus the high-resolution SAR images of complex moving vessels in high sea states. According to the characteristics of the high-order multi-component polynomial phase signal, this algorithm provides parallel processing and co-evolution methods by setting the different permissions of the sub-population and sharing its search information. As a result, the multiple components can be extracted simultaneously. Experiments were conducted using the simulation data and Gaofen-3 (GF-3) SAR data. Results showed the processing speed increased by more than 40% and the global convergence was significantly improved. The imaging results verify the efficiency and robustness of this co-evolutionary PSO.

Modeling River Discharge Using Automated River Width Measurements Derived from Sentinel-1 Time Series

Against the background of a worldwide decrease in the number of gauging stations, the estimation of river discharge using spaceborne data is crucial for hydrological research, river monitoring, and water resource management. Based on the at-many-stations hydraulic geometry (AMHG) concept, a novel approach is introduced for estimating river discharge using Sentinel-1 time series within an automated workflow. By using a novel decile thresholding method, no a priori knowledge of the AMHG function or proxy is used, as proposed in previous literature. With a relative root mean square error (RRMSE) of 19.5% for the whole period and a RRMSE of 15.8% considering only dry seasons, our method is a significant improvement relative to the optimized AMHG method, achieving 38.5% and 34.5%, respectively. As the novel approach is embedded into an automated workflow, it enables a global application for river discharge estimation using solely remote sensing data. Starting with the mapping of river reaches, which have large differences in river width over the year, continuous river width time series are created using high-resolution and weather-independent SAR imaging. It is applied on a 28 km long section of the Mekong River near Vientiane, Laos, for the period from 2015 to 2018.

SAR Image Ship Object Generation and Classification With Improved Residual Conditional Generative Adversarial Network

Image generation by the conditional generative adversarial network (CGAN) can provide rich samples for supervised classification tasks. However, the unstable gradient updating problem makes it difficult to be used for image generation, especially for the high-resolution SAR image. In this letter, an improved residual condition generation network is proposed to enhance the quality of generated images and classification accuracy for hard negative examples. First, a residual convolutional-based block is built to clear the detail texture of different types of targets. Then, the gradient penalty and Wasserstein loss are used as discriminators to improve the similarity between real samples and the intra diversity of a generated images. We test the proposed method on generation and classification of three kinds of commercial ships using C-band 3-m Gaofen-3 (GF-3) SAR images. Experimental results show that our method can generate high-quality ship samples and achieve good classification accuracy.

Evaluation of passive microwave melt detection methods on Antarctic Peninsula ice shelves using time series of Sentinel-1 SAR

Passive microwave datasets have been used to quantify the extent and duration of surface melt in Greenland and Antarctica from 1978 on with daily and near-daily intervals. These results have important implications for climate analysis and may help evaluate ice shelf stability. However, the accuracy of passive microwave methods used to detect melt is difficult to quantify, especially on the Antarctic Peninsula. Here four different melt detection methods are employed, including a new formulation of a statistical analysis of brightness temperature time series using a K-means clustering algorithm. Strikingly, two of the most widely used passive microwave melt detection methods are found to vary by 48% mean days of melt per year across six different locations on the Larsen C, Wilkins, and George VI Ice Shelves. In the absence of ground truth observations, time series of Sentinel-1 SAR observations from 2016 on provide a comparison dataset. In topographically flat regions where surface melt is spatially uniform, the passive microwave melt detection method based on a K-means analysis and the cross-polarization gradient ratio method demonstrate the highest agreement and correlation with active radar melt detection methods. One issue which has plagued passive microwave analysis is its coarse spatial resolution. High resolution SAR images are able to demonstrate and quantify the spatial variability of melt within individual passive microwave pixels. Melt is shown to be suitably uniform in space for passive microwave applications at the study sites on Antarctic Peninsula ice shelves, but not so in other regions of the Antarctic Peninsula. Spatial heterogeneity of surface melt on the sub-pixel scale is often related to varying surface topography.

Inversion and Analysis of Mining Subsidence by Integrating DInSAR, Offset Tracking, and PIM Technology

High-intensity underground mining generates considerable surface subsidence in mining areas, including ground cracks and collapse pits on roads and farmland, threatening the safety of buildings. Large-amplitude subsidence (e.g., >2 m) is usually characterized by a large phase gradient in interferograms, leading to severe phase decorrelation and unwrapping errors. Therefore, the subsidence on the surface cannot be well derived simply using conventional differential interferometric synthetic aperture radar (DInSAR) or other geodetic measurements. We propose a new method that combines both DInSAR and subpixel offset-tracking technology to improve mine subsidence monitoring over large areas. We utilize their respective advantages to extract both the spatial boundaries and the amplitude of displacements. Using high-resolution RADARSAT-2 SAR images (5 m) acquired on February 13, 2012, and November 27, 2012, in the Shendong Coalfield located at the border between Shaanxi Province and Inner Mongolia Province, China, we obtain the subcentimetre-level subsidence of the mine boundary by DInSAR and resolve the metre-level mine subsidence centre based on subpixel offset tracking. The whole subsidence field is obtained by combining and analyzing the subcentimetre-level and the metre-level subsidence. We use the probability integral method (PIM) function model to fit the boundary and central mine subsidence to reconstruct the spatial distribution of the mine subsidence. Our results show that the maximum central subsidence reaches ~4.0 m (beyond the monitoring capabilities of DInSAR), which is generally in agreement with the maximum subsidence of ~4.0-5.0 m from field investigation. We also model the boundary and the central subsidence (the final fitting coefficient is 0.978). Our findings indicate that the offset-tracking method can compensate for the deficiency of DInSAR in large-amplitude subsidence extraction, and the inclusion of the PIM technique helps reconstruct the whole subsidence field in mining areas.

Object-based landfast sea ice detection over West Antarctica using time series ALOS PALSAR data

Landfast sea ice (fast ice) is an important feature prevalent around the Antarctic coast, which is affected by climate change and energy exchanges with the atmosphere and ocean. This study proposed a method for detection of the West Antarctic fast ice using the Advanced Land Observing Satellite Phased Array L-band SAR (ALOS PALSAR) images. The algorithm has combined image segmentation, image correlation analysis, and machine learning techniques (i.e., random forest (RF), extremely randomized trees (ERT), and logistic regression (LR)). We used SAR images with a baseline of 5 days that are not in the same orbit but overlap each other as overlaps between swaths in adjacent orbits are often available in the polar regions. The underlying assumption for the proposed fast ice detection algorithm is that fast ice regions in SAR images with a time interval of 5 days are highly correlated. The object-based approach proposed in this study was well suited to high-resolution SAR images in deriving spatially homogeneous fast ice regions. The image segmentation results using the optimized parameters showed a distinct difference in the backscatter temporal evolution between fast ice and pack ice regions. Correlation and STD of backscattering coefficients were found to be the most significant variables for the object-based fast ice detection from two temporally separated images. In overall, the quantitative and qualitative evaluation demonstrated that the algorithm was an effective approach to detect fast ice with high accuracies. The models well detected various fast ice regions in the West Antarctica but misclassified some objects. The misclassifications occurred toward the edge of fast ice regions with relatively rapid changes in backscattering between both data acquisitions. On the other hand, few fast ice objects were misclassified as uniform backscattering over time occurred by chance on very small objects far from the coast. Very old multi-year fast ice regions with high backscattered signals were also a source for some misclassifications. This may be due to the sensitivity of L-band to snow structure to some extent and a thinner ice over the region with either ice growth (no deformation) or closing (slight deformation) between both images. Heavy snow load on the ice could be another error source for some misclassification as well. The approach allowed for the reliable detection of fast ice regions by using L-band SAR images with a small local incidence angle difference.

Monitoring Rock Glacier Kinematics with Satellite Synthetic Aperture Radar

Active rock glaciers represent the best visual expression of mountain permafrost that can be mapped and monitored directly using remotely sensed data. Active rock glaciers are bodies that consist of a perennially frozen ice/rock mixture and express a distinct flow-like morphology indicating downslope permafrost creep movement. Annual rates of motion have ranged from a few millimeters to several meters per year, varying within the annual cycle, from year to year, as well as at the decennial time scale. During the last decade, in situ observations in the European Alps have shown that active rock glaciers are responding almost synchronously to inter-annual and decennial changes in ground temperature, suggesting that the relative changes of their kinematics are a general indicator of the evolution of mountain permafrost conditions. Here, we used satellite radar interferometry (InSAR) to monitor the rate of motion of various active rock glaciers in the Swiss Alps, Qeqertarsuaq (Western Greenland), and the semiarid Andes of South America. Velocity time series computed with Sentinel-1 SAR images, regularly acquired since 2014, every six days over Europe and Greenland and every 12 days over the Andes, show annual fluctuations, with higher velocities at the end of the summer. A JERS-1 image pair of 1996 and stacks of very high-resolution SAR images from TerraSAR-X and Cosmo-SkyMed from 2008 to 2017 were analyzed using InSAR and offset tracking over the Western Swiss Alps in order to extend the main observation period of our study. A quantitative assessment of the accuracy of InSAR and offset tracking was performed by comparison with in situ methods. Our results for the three different study regions demonstrate that Sentinel-1 InSAR can complement worldwide in situ measurements of active rock glacier kinematics.

Classification of Large-Scale High-Resolution SAR Images With Deep Transfer Learning

The classification of large-scale high-resolution SAR land cover images acquired by satellites is a challenging task, facing several difficulties such as semantic annotation with expertise, changing data characteristics due to varying imaging parameters or regional target area differences, and complex scattering mechanisms being different from optical imaging. Given a large-scale SAR land cover dataset collected from TerraSAR-X images with a hierarchical three-level annotation of 150 categories and comprising more than 100,000 patches, three main challenges in automatically interpreting SAR images of highly imbalanced classes, geographic diversity, and label noise are addressed. In this letter, a deep transfer learning method is proposed based on a similarly annotated optical land cover dataset (NWPU-RESISC45). Besides, a top-2 smooth loss function with cost-sensitive parameters was introduced to tackle the label noise and imbalanced classes' problems. The proposed method shows high efficiency in transferring information from a similarly annotated remote sensing dataset, a robust performance on highly imbalanced classes, and is alleviating the over-fitting problem caused by label noise. What's more, the learned deep model has a good generalization for other SAR-specific tasks, such as MSTAR target recognition with a state-of-the-art classification accuracy of 99.46%.

Electric Tower Target Identification Based on High-resolution SAR Image and Deep Learning

With regard to the electric tower identification problem in the high-resolution SAR image under complicated background, it is difficult for the existing target detection algorithm to reach balance in identification precision and efficiency. Therefore, this paper combines the advantages of high YOLOv2 calculation efficiency and high VGG classification precision, and proposes a kind of two-stage target detection algorithm of YOLOv2 and VGG cascade connection by virtue of transfer learning. At the Stage-1, the sliding window and the non-maximum suppression algorithms are combined, and the YOLOv2 is utilized to conduct rapid electric power detection to the whole SAR image, thus obtaining the target detection result of high recall rate; at the Stage-2, the VGG classification model is utilized to conduct secondary classification to the target and background with regard to the target detection result at the Stag-1, thus further eliminating the false positive. This algorithm can be used to enhance the accuracy of only using YOLO v2 f or target identification, thus reducing the false alarm rate of model effectively. The electric towers in the mountainous belt and the plain belt are used as samples for training, thus enhancing the robustness of algorithm. After conducting algorithm testing to the COSMO image in Zhijiang City, Hubei Province, the result shows that the electric tower recall rate can reach 73.8%. The method adopted in this paper can identify the electric tower target of the whole SAR image more accurately through model transfer and two-stage deep learning, thus being further promoted into actual application scene.

High Resolution SAR Image Algorithm with Sample Length Constraints for the Range Direction

The traditional Range Doppler (RD) algorithm is unable to meet practical needs owing to the limit of resolution. The order of fractional Fourier Transform (FrFT) and the length of sampling signals affect SAR imaging perfo... | Find, read and cite all the research you need on Tech Science Press

Deformation Feature Analysis of Qinghai–Tibet Railway Using TerraSAR-X and Sentinel-1A Time-Series Interferometry

The Qinghai–Tibet Railway (QTR) is the highest plateau artificial facility, connecting Lhasa and Golmud over Qinghai–Tibet Plateau. Climate change and anthropogenic activities are changing the condition of plateau, with potential influences on the stabilities of QTR. Synthetic aperture radar interferometry (InSAR) technique could retrieve ground millimeter scale deformation utilizing phase information from SAR images. In this study, the structure and deformation features of QTR are retrieved and analyzed using time-series interferometry with Sentinel-1A and TerraSAR-X images. The backscattering and coherence features of QTR are analyzed in medium and very high-resolution SAR images. Then, the deformation results from different SAR datasets are estimated and analyzed. Experimental results show that some of the QTR sections undergo serious deformation, with the maximum deformation rate of –20 mm/year. Moreover, the detailed deformation feature in the Beiluhe section has been analyzed as well as the effects of different cooling measurements underline QTR embankment. It is also found that embankment-bridge transition section along QTR is prone to undergo deformation. Our study demonstrates the application potential of high-resolution InSAR in deformation monitoring of QTR.

Sparse Scene Recovery for High-Resolution Automobile FMCW SAR via Scaled Compressed Sensing

This paper introduces a sparse scene reconstruction algorithm for automobile frequency-modulated continuous-wave synthetic aperture radar (FMCW SAR) through scaled compressed sensing (CS). An FMCW radar leads to low manufacturing cost, compact realization, and low transmission power. An automobile SAR is more economical and easier to implement than typical SAR platforms (e.g., satellites and aircraft). We apply CS to randomly subsampled raw data of automobile FMCW SAR for sparse reconstruction. We exploit the fact that the velocity of an automobile is significantly lower than that of the SAR platforms, which leads to a vastly narrow bandwidth of an azimuth-matched filter in the azimuth compression of the range-Doppler algorithm (RDA). Low-frequency azimuth data have a fundamental effect on azimuth compression. We propose a new reconstruction scheme, scaled CS, which specializes in low-frequency information recovery for automobile SAR. The scheme is based on basis pursuit denoising (BPDN). A Ku-band FMCW SAR system is developed to verify the performance of the proposed algorithm. We mount our system on an automobile and collect FMCW SAR raw data in the stripmap mode with a van maintained a constant speed on a highway. The proposed reconstruction algorithm shows improved recovery performance for automobile FMCW SAR, which is validated by processing a high-resolution real SAR image.

A 3D building reconstruction method for SAR images based on deep neural network

Owing to its momentous applications in urban planning, disaster monitoring, smart cities and other fields, 3D building reconstruction is an important topic in many research areas such as computer vision, photogrammetry, and remote sensing. The particularity and complexity of the microwave scattering mechanism bring great challenges to the 3D building reconstruction of SAR images, and the applicability and automation of existing methods need to be improved. This study constructs the overall framework of building detection in SAR images and 3D reconstruction based on deep learning and radar imaging mechanism. It puts forward a method of using a coupled equivalent complex valued convolutional neural network for building facade detection in SAR images, a method for RaySAR-based modeling simulation and point cloud generation for 3D model training, and a method for a 3D generation network for 3D building reconstruction from SAR images. Experiments using TerraSAR-X and GF-3 high resolution SAR images are carried out, producing good 3D reconstruction results. The proposed method is a useful attempt for 3D target reconstruction of SAR images and represents a new technical approach.

USING COHERENT SCATTERERS IN TIME SERIES OF HIGH RESOLUTION SAR IMAGES FOR THE MONITORING OF CONSTRUCTION ACTIVITY

Abstract. In this paper, we present a method for detecting and quantifying changes due to construction activity in time series of high resolution SAR images. We propose to detect these changes by the appearance or disappearance of a large number of coherent scatterers, rather than by looking at changes in SAR amplitude or coherence. These coherent scatterers, which appear mainly in buildings and other man-made structures, will be detected in each individual image by analysing their stability across multiple frequency sub-bands. This implies that long time series are not required, unlike for other similar approaches based in persistent scatterer interferometry, and that this approach can be applied for the monitoring of fast changing areas such as temporary settlements. The presented approach will be applied for monitoring the construction activity taking place during the preparations for two large festivals in Munich, Germany.

Geostatistical Integration of Field Measurements and Multi-Sensor Remote Sensing Images for Spatial Prediction of Grain Size of Intertidal Surface Sediments

Park, N.-W., 2019. Geostatistical integration of field measurements and multi-sensor remote sensing images for spatial prediction of grain size of intertidal surface sediments. In: Jung, H.-S.; Lee, S.; Ryu, J.-H., and Cui, T. (eds.), Advances in Remote Sensing and Geoscience Information Systems of Coastal Environments. Journal of Coastal Research, Special Issue No. 90, pp. 190-196. Coconut Creek (Florida), ISSN 0749-0208.The objective of this paper is to demonstrate the potential benefit of using high-resolution optical and SAR images for the geostatistical mapping of grain size of intertidal surface sediments. The grain size values from field measurements are integrated with reflectance and backscattering coefficients from multi-sensor images via regression kriging. The trends of grain size variations are estimated using support vector regression (SVR) modeling to account for a nonlinear relationship, and rank transformation is applied to original input variables to highlight the relative differences in input values from multi-sensor images. Unlike the conventional regression-based mapping approach, the residual component that cannot be explained by the multi-sensor remote sensing images is considered and predicted via kriging. The final grain size values are then obtained by adding these two components. From a case study on the Baramarae tidal flats in Korea with KOMPSAT-2 and COSMO-SkyMed images, the integration of multi-sensor images with field measurements via SVR and rank transformation could explain 58 % of grain size variance, leading to a significant improvement in predictive performance (approximately 29 %) over ordinary kriging based on field measurements only. Furthermore, using reflectance and scattering information from multi-sensor images generated the grain size distribution with more detailed variations in the study area. Therefore, the synergistic use of multi-sensor images within an advanced geostatistical integration framework is expected to be very effective for the reliable mapping of the grain size of intertidal surface sediments when only a limited number of field measurements are available.

High-Rise Building 3D Reconstruction with the Wrapped Interferometric Phase.

The great development of high-resolution SAR system gives more opportunities to observe building structures in detail, especially the advanced interferometric SAR (InSAR), which techniques attract more attention on exploiting useful information on urban infrastructures. Considering that the high-rise buildings in urban areas are quite common in big cities, it is of great importance to retrieve the three-dimension (3D) information of the urban high-rise buildings in urban remote sensing applications. In this paper, the 3D reconstruction of high-rise buildings using the wrapped InSAR phase image was studied, referring to the geometric modulation in very high resolution (VHR) SAR images, such as serious layover cause by high-rise buildings. Under the assumption of a rectangular shape, the high-rise buildings were detected and building façades were extracted based on the local frequency analysis of the layover fringe patterns. Then 3D information of buildings were finally extracted according to the detected façade geometry. Except for testing on a small urban area from the TanDEM-X data, the experiment carried on the single-pass InSAR wrapped phase in the wide urban scene, which was collected by the Chinese airborne N-SAR system, also demonstrated the possibility and applicability of the approach.

Resolution Analysis of Circular Synthetic Aperture Radar Noncoherent Imaging

Circular synthetic aperture radar (CSAR) has attracted much attention in the field of high-resolution SAR imaging. The spatial resolution is usually considered as the most significant parameter for SAR image quality evaluation and has been widely applied for different monostatic/bistatic SAR operation modes. However, it is hard to get an analytical solution to calculate exactly the value of spatial resolution, especially for the circumstance of CSAR noncoherent imaging. In this paper, based on general impulse response function (IRF) of CSAR noncoherent imaging, a method to obtain the quasi-analytical result of the spatial resolution is proposed, which takes the fractional bandwidth and the subaperture angle into consideration. First, the IRF for CSAR is introduced. Second, based on the IRF, the spatial resolution of CSAR noncoherent imaging is derived. Moreover, the half-power beam-width factor is presented, which allows obtaining the form of spatial resolution in an easier way. Finally, the effectiveness of the proposed method is verified with the simulated and experimental CSAR data.

Flood depth estimation by means of high-resolution SAR images and lidar data

Abstract. When floods hit inhabited areas, great losses are usually registered in terms of both impacts on people (i.e., fatalities and injuries) and economic impacts on urban areas, commercial and productive sites, infrastructures, and agriculture. To properly assess these, several parameters are needed, among which flood depth is one of the most important as it governs the models used to compute damages in economic terms. This paper presents a simple yet effective semiautomatic approach for deriving very precise inundation depth. First, precise flood extent is derived employing a change detection approach based on the normalized difference flood index computed from high-resolution synthetic aperture radar imagery. Second, by means of a high-resolution lidar digital elevation model, water surface elevation is estimated through a statistical analysis of terrain elevation along the boundary lines of the identified flooded areas. Experimental results and quality assessment are given for the flood that occurred in the Veneto region, northeastern Italy, in 2010. In particular, the method proved fast and robust and, compared to hydrodynamic models, it requires sensibly less input information.

Automatic Ship Detection Using the Artificial Neural Network and Support Vector Machine from X-Band Sar Satellite Images

In this paper, an automatic ship detection method using the artificial neural network (ANN) and support vector machine (SVM) from X-band SAR satellite images is proposed. When using machine learning techniques, the most important points to consider are (i) defining the proper input neurons and (ii) selecting the correct training data. We focused on generating two optimal input data neurons that (i) strengthened ship targets and (ii) mitigated noise effects by image processing techniques, including median filtering, multi-looking, etc. The median filter and multi-look operations were used to reduce the background noise, and the median filter operation was also used to remove ships in an image in order to maximize the difference between the pixel values of ships and the sea. Through the root-mean-square difference calculation, most ship targets, even including small ships, were emphasized in the images. We tested the performance of the proposed method using X-band high-resolution SAR images including COSMO-SkyMed, KOMPSAT-5, and TerraSAR-X images. An intensity difference map and a texture difference map were extracted from the X-band SAR single-look complex (SLC) images, and then, the maps were used as input neurons for the ANN and SVM machine learning techniques. Finally, we created ship-probability maps through the machine learning techniques. To validate the ANN and SVM results, optimal threshold values were obtained by using the statistical approach and then used to identify ships from the ship-probability maps. Consequently, the level of recall achieved was greater than 90% in most cases. This means that the proposed method enables the detection of most ship targets from X-band SAR images with a reduced number of false detections from negative effects.