Unifying Experiment Design and Convex Regularization Techniques for Enhanced Imaging With Uncertain Remote Sensing Data—Part I: Theory

This paper considers the problem of high-resolution imaging of the environment formalized in terms of a nonlinear ill-posed inverse problem of nonparametric estimation of the power spatial spectrum pattern (SSP) of the wavefield scattered from an extended remotely sensed scene (referred to as the scene image) via processing the discrete measurements of a finite number of independent realizations of the observed degraded radar data signals (one realization of the trajectory signal in the case of SAR). The model-level uncertainties are associated with unknown statistics of perturbations of the signal formation operator (SFO) in turbulent environment. The system-level uncertainties are attributed to the imperfect array calibration, finite dimensionality of measurements, uncontrolled antenna vibrations and random carrier trajectory deviations in the case of SAR. An effective method for SSP reconstruction is therefore proposed by incorporating into the minimum risk (MR) nonparametric spectral estimation strategy the experiment design-motivated constraints of SSP observability/identifiability for the finite-dimensional range continuous-to-discrete SFO algorithmically coupled with descriptive experiment design regularization (DEDR) and unified with worst-case statistical performance optimization approach. The MR objective functional is constrained by this information, and the robust DEDR reconstruction operator applicable to the scenarios with the low-rank uncertain estimated data correlation matrices is found. We also show how this algorithm may be considered generalization of the robust MVDR and the regularized inverse spatial filtering techniques. The efficiency of the developed technique is illustrated via numerical simulations.

We address a new approach for enhanced microwave remote sensing (RS) imaging via performing the imaging system kernel point spread function (PSF) operator refinement-based multi-scale iterative reconstructive (MSIR) image post-processing, as required for emerging feature enhanced RS missions. The high-resolution (HR) image is first reconstructed from the initial low-resolution (LR) image employing the statistically optimal minimum risk inspired descriptive experiment design regularization (DEDR) framework. Next, to approach the overall super-resolution (SR) imaging performances we incorporate into the DEDR method the additional postprocessing stage aimed at filling in the null space of the HR imaging system PSF operator via performing the corresponding projections onto the nested refined resolution greed frames. We feature the differences between the proposed DEDR-MSIR resolution refinement approach and the most competing celebrated Papoullis-Gerchberg SR method adapted for the feature enhanced RS imaging and demonstrate the advantages of the unified DEDR-MSIR approach for SR image recovery.

The convex optimization-based descriptive experiment design regularization (DEDR) method is aggregated with the neural network (NN)-adapted variational analysis (VA) approach for adaptive high-resolution sensing into a unified DEDR -VA-NN framework that puts in a single optimization frame high-resolution radar/SAR image formation in uncertain operational scenarios, adaptive despeckling and dynamic scene image enhancement for a variety of sensing modes. The DEDR -VA-NN method outperforms the existing adaptive radar imaging techniques both in resolution and convergence rate. The simulation examples are incorporated to illustrate the efficiency of the proposed DEDR-VA-related imaging techniques.

We address a new approach for computational enhancement of low resolution radar/fractional SAR imagery that unifies the descriptive experiment design regularization (DEDR) framework with the total variation (TV) inspired sparsity preserving (Sp-Pr) regularization paradigm. The new framework incorporates the structured projectors onto the convex solution sets (POCS) into the TV-restructured DEDR framework and solves the overall compressed image computational reconstruction/enhancement problem employing the proposed aggregated DEDR-TV-POCS strategy.

A new aggregated Hardware/Software (HW/SW) codesign approach to optimization of the digital signal processing techniques for enhanced imaging with real-world uncertain remote sensing (RS) data based on the concept of descriptive experiment design regularization (DEDR) is addressed. We consider the applications of the developed approach to typical single-look synthetic aperture radar (SAR) imaging systems operating in the real-world uncertain RS scenarios. The software design is aimed at the algorithmic-level decrease of the computational load of the large-scale SAR image enhancement tasks. The innovative algorithmic idea is to incorporate into the DEDR-optimized fixed-point iterative reconstruction/enhancement procedure the convex convergence enforcement regularization via constructing the proper multilevel projections onto convex sets (POCS) in the solution domain. The hardware design is performed via systolic array computing based on a Xilinx Field Programmable Gate Array (FPGA) XC4VSX35-10ff668 and is aimed at implementing the unified DEDR-POCS image enhancement/reconstruction procedures in a computationally efficient multi-level parallel fashion that meets the (near) real-time image processing requirements. Finally, we comment on the simulation results indicative of the significantly increased performance efficiency both in resolution enhancement and in computational complexity reduction metrics gained with the proposed aggregated HW/SW co-design approach.

We address a new multimode system/method fusion oriented neural network (NN) computing approach to enhancement of conventional low resolution remote sensing (RS) radar and/or fractional synthetic aperture radar imagery. First, the squared error norm objective function minimization-based descriptive experiment design regularization (DEDR) framework is adapted to the Hopfield-type neural network computing-based feature enhancing image reconstruction from the low resolution initial RS imagery. Second, the DEDR framework is aggregated with the variational analysis inspired total variation (TV) minimization modality aimed at anisotropic feature-enhanced image recovery with locally selective information fusion adaptively balanced over speckle and noise suppression. The DEDR and the TV enhancement modalities are fused into the TV-structured maximum entropy neural network (MENN) technique. The developed DEDR-TV-structured MENN-implemented RS image enhancement method outperforms the recently proposed competing approaches both in the achievable resolution enhancement over noise suppression and the convergence rates that is corroborated via the reported simulations.

We address a new approach to solving radar imaging problems stated and treated as uncertain ill-conditioned inverse problems of nonparametric spatial power spectrum estimation via processing the finite number of independent observations of the degraded array data signals (one realization of the trajectory signal in the case of SAR). The idea is to adapt a statistically optimal minimum risk nonparametric power spectrum estimation approach to the radar imaging scenarios with model-level and system-level uncertainties. The proposed incorporation of the worst-case performance optimization-adapted robust regularization aggregated with the descriptive experiment design paradigm into the minimum risk nonparametric estimation strategy leads to a new unified doubly regularized minimum risk approach for robust adaptive high-resolution reconstructive imaging in the uncertain remote sensing scenarios.

A high-speed dual super-systolic core for reconstructive signal processing (SP) operations consists of a double parallel systolic array (SA) machine in which each processing element of the array is also conceptualized as another SA in a bit-level fashion. In this study, we addressed the design of a high-speed dual super-systolic array (SSA) core for the enhancement/reconstruction of remote sensing (RS) imaging of radar/synthetic aperture radar (SAR) sensor systems. The selected reconstructive SP algorithms are efficiently transformed in their parallel representation and then, they are mapped into an efficient high performance embedded computing (HPEC) architecture in reconfigurable Xilinx field programmable gate array (FPGA) platforms. As an implementation test case, the proposed approach was aggregated in a HW/SW co-design scheme in order to solve the nonlinear ill-posed inverse problem of nonparametric estimation of the power spatial spectrum pattern (SSP) from a remotely sensed scene. We show how such dual SSA core, drastically reduces the computational load of complex RS regularization techniques achieving the required real-time operational mode.

Sufficient remote sensing observation data during crop main growing season is of great importance in improving the accuracy of data assimilation of crop model. The optical remote sensing data are susceptible to cloud and rain, so the amount of clear optical data is very limited in cloudy weather or rainy day. Synthetic Aperture Radar (SAR) is not dependent on cloud cover or light conditions, it can penetrate through clouds and have all-weather capabilities. This allows for a more reliable and consistent crop monitoring and yield estimation in terms of radar sensor data. So, the aim of this article is to improve the accuracy for winter wheat yield estimation by joint assimilation of SAR and optical satellite images into crop model. In this study, SAR images are acquired by C-band SAR sensor boarded on Sentinel-1 satellites, and optical images are obtained from Sentinel-2 satellites. Remote sensing data and ground data are all collected during the main growth and development stages of winter wheat. Both normalized difference vegetation index (NDVI) derived from Sentinel-2 images and backscattering coefficients and polarimetric indicators computed from Sentinel-1 images are used in water cloud model to derive soil moisture (SM) time series images. To improve the prediction of crop yields at filed scale, we incorporate remotely sensed soil moisture into the WOrld FOod STudies (WOFOST) model using Ensemble Kalman filter (EnKF) algorithm. In general, the results show that data assimilation schemes of remotely sensed soil moisture slightly improved the correlation of observed and simulated yields (R2 = 0.30; RMSE =782 kg ha−1) compared to the situation without data assimilation (R2 = 0.14; RMSE = 1398 kg ha−1). Results of this study indicate that the potential for assimilation SAR and optical remote sensing data to improve field yield estimates is relatively low, limitations are due to insufficient no-cloud optical remotely sensed data and root zone soil moisture information. Consequently, the results of this study demonstrate the potential and usefulness of assimilating both SAR and optical remote sensing data into crop model for crop monitoring and yield estimation. Moreover, this also provides the reference for crop yield estimation with data assimilation in other agricultural landscapes.

An interesting relationship between the probability-constrained and worst-case optimization based robust minimum variance (MV) beamformers has been discovered. It is shown that both in the cases of circularly symmetric Gaussian and worst-case distributions of the steering vector mismatch, the probability-constrained robust MV beamforming problem can be tightly approximated as a convex second-order cone programming (SOCP) problem. The latter problem is mathematically equivalent to that resulting from the deterministic worst-case approach and, therefore, probability-constrained beamformers can be interpreted and implemented using their deterministic worst-case counterparts. However, an important advantage of the developed probability-constrained MV beamformers with respect to their standard worst-case counterparts is that the former approaches enable to explicitly quantify the parameters of the uncertainty region in terms of the beamformer outage probability.

Archaeological prospecting is pivotal to understanding human civilization’s development and past human–environment interactions (PHEIs). Synthetic aperture radar (SAR), a unique active remote sensing (RS) technology capable of penetrating vegetation and dry soil, has become an indispensable tool for large-scale archaeological surveys. This review, encompassing 45 years of progress (1980–2024), systematically chronicles SAR’s evolution—a field we term <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SARchaeology</i>—from pioneering L-band discoveries to sophisticated modern techniques like high-resolution SAR (HiSAR), interferometric SAR (InSAR), polarimetric SAR (PolSAR), and multitemporal SAR analysis (MtSAR). These methods have enabled unprecedented discoveries of hidden landscapes, particularly in areas obscured by dense cover or aridity. However, SARchaeology grapples with inherent limitations we define as the “SAR gap,” driven by technical complexity and, critically, the difficulty in achieving adequate signal-to-noise ratios (SNR) for small low-backscatter archaeological targets. Traditional challenges like speckle noise, cost, and interpretation ambiguity are compounded by this fundamental issue. To bridge this gap, we propose an integrated framework of five strategic frontiers: 1) leveraging innovative long-wavelength SAR technologies (e.g., the P band) for deep penetration; 2) establishing collaborative space–air–ground networks; 3) integrating SAR with complementary RS techniques; 4) implementing artificial intelligence (AI) and machine learning (ML) for noise suppression and weak signal feature extraction; and 5) developing explainable SARchaeology (X-SARchaeology). This framework provides a definitive road map for the systematic exploration of buried civilizations in hyper-arid regions (e.g., the Silk Road and ancient Mesopotamia). By focusing on these solutions, SARchaeology is poised to transcend conventional limitations and illuminate much more of human history before its Golden Jubilee.

High-spatiotemporal-resolution and accurate soil moisture (SM) data are crucial for investigating climate, hydrology, and agriculture. Existing SM products do not yet meet the demands for high spatiotemporal resolution. The objective is to develop and evaluate a retrieval framework to derive SM estimates with high spatial (100 m) and temporal (&lt;3 days) resolution that can be used on a national scale in China. Therefore, this study integrates multi-source data, including optical remote sensing (RS) data from Sentinel-2 and Landsat-7/8/9, synthetic aperture radar (SAR) data from Sentinel-1, and auxiliary data. Four machine learning and deep learning algorithms are applied, including Random Forest Regression (RFR), Extreme Gradient Boosting (XGBoost), Long Short-Term Memory (LSTM) networks, and Ensemble Learning (EL). The integrated framework (IF) considers three feature scenarios (SC1: optical RS + auxiliary data, SC2: SAR + auxiliary data, SC3: optical RS + SAR + auxiliary data), encompassing a total of 33 features. The results are as follows: (1) The correlation coefficients (r) between auxiliary data (such as sand fraction, r = −0.48; silt fraction, r = 0.47; and evapotranspiration, r = −0.42), SAR features (such as the backscatter coefficients for VV-pol (σvv0), r = 0.47), and optical RS features (such as Shortwave Infrared Band 2 (SWIR2) reflectance data from Sentinel-2 and Landsat-7/8/9, r = −0.39) with observed SM are significant. This indicates that multi-source data can provide complementary information for SM monitoring. (2) Compared to XGBoost and LSTM, RFR and EL demonstrate superior overall performance and are the preferred models for SM prediction. Their R2 for the training and test sets exceed 0.969 and 0.743, respectively, and their ubRMSE are below 0.022 and 0.063 m3/m3, respectively. (3) The SM prediction accuracy is highest for the scenario of optical + SAR + auxiliary data, followed by SAR + auxiliary data, and finally optical + auxiliary data. (4) With an increasing Normalized Difference Vegetation Index (NDVI) and SM values, the trained models exhibit a general decrease in prediction performance and accuracy. (5) In 2021 and 2022, without considering cloud cover, the IF theoretically achieved an SM revisit time of 1–3 days across 95.01% and 96.53% of China’s area, respectively. However, SC1 was able to achieve a revisit time of 1–3 days over 60.73% of China’s area in 2021 and 69.36% in 2022, while the area covered by SC2 and SC3 at this revisit time accounted for less than 1% of China’s total area. This study validates the effectiveness of combining multi-source RS data with auxiliary data in large-scale SM monitoring and provides new methods for improving SM retrieval accuracy and spatiotemporal coverage.

Corn is an important food crop worldwide, and its yield is directly related to Chinese food security. Accurate remote sensing extraction of corn can realize the rational application of land resources, which is of great significance to the sustainable development of modern agriculture. In the field of large-scale crop remote sensing classification, single-period optical remote sensing images often cannot achieve high-precision classification. To improve classification accuracy, multiple time series image combinations have gradually been adopted. However, due to the influence of cloudy and rainy weather, it is often difficult to obtain complete time series optical images. Synthetic aperture radar (SAR) data are imaged by microwaves, which have strong penetrating power and are not affected by clouds. A critical way to solve this problem is to use SAR images to compensate for the lack of optical images and obtain a complete time series image in the corn-growing season. However, SAR images have limited wavelengths and cannot provide important wavelengths, such as visible light bands and near-infrared information. To solve this problem, this study took Zhaodong City, a vital corn-planting base in China, as the research area; took GF-6/GF-3 and Sentinel-1/Sentinel-2 as remote sensing data sources; designed12 classification scenarios; analyzed the best classification period and the best time series combination of corn classification; studied the influence of SAR images on the classification results of time series images; and compared the classification differences between GF-6/GF-3 and Sentinel-1/Sentinel-2. The results show that the classification accuracy of time series combinations is much higher than that of single-period images. The polarization characteristics of SAR images can improve the classification accuracy with time series images. The classification accuracy of GF series images from China is obviously higher than that of Sentinel series images. The research performed in this paper can provide a reference for agricultural classification by using remote sensing data.

Rock glaciers represent a typical landform characterizing high mountain periglacial terrains. They are composed by a mixture of frozen debris and interstitial ice and when are affected by downslope displacement related to the permafrost deformation or melting, they are called active rock glaciers. The observation of their changes over time is a critical indicator of the state of water resources and permafrost distribution and is therefore of great importance for risk scenario definition and related natural hazards management, as well as can be related to the climate change.&amp;#160; Synthetic Aperture Radar (SAR) data have proven to be a useful tool to estimate surface displacement phenomena in mountainous areas that can be difficult to access. In particular, Differential Interferometry can measure displacement in the line of site direction with high accuracy. However, sometimes the characteristics of the areas in terms of displacement rate and number of available images lead to some limitations, which can be overcome by using other complementary tools such as SAR offset-tracking (OT) or by integrating both techniques together. In the OT case, the estimation of the deformation in azimuth and range directions is obtained maximizing the normalized cross-correlation between pairs of images and the minimum detectable displacement depends on the spatial resolution of SAR images, in the sense that a finer spatial resolution allows estimating a lower minimum detectable displacement. Consequently, these techniques can be profitably used for medium- or long-term displacement measurements [1]. In this work, OT is used to monitor the active alpine Lazaun rock glacier, located in the &amp;#214;tztal in South Tyrol (Italy). Lazaun has a snow cover free period limited typically to 3 months per year, covers an area of about 0.12 km&amp;#178;, and its altitude is between 2480 and 2700 m a.s.l.. Previous measurements performed by using GPS have detected displacements reaching ca. 1.5 m per year [2]. Different kind of data (TerraSAR-X, Cosmo-Skymed First and Second generation, Sentinel-1 and SAOCOM) characterized by different wavelengths, exposure and resolutions have been tested to estimate the rock glacier displacements and the results have been compared with GPS measurements executed in 2016-2018 and 2022. The results show that considering the limited size of the area of interest and the displacement rate, the spatial resolution of the data is of fundamental importance and only using spotlight SAR data is possible to estimate the displacement on both seasonal and inter-annual temporal scale. This research is part of the 2021-2023 project &amp;#8216;CRIOSAR: Applicazioni SAR multifrequenza alla criosfera&amp;#8217;, funded by ASI under grant agreement n. 2018-12-U.0. TerraSAR-X data were provided by the European Space Agency, Project Proposal id 34722, &amp;#169; DLR, distribution Airbus DS Geo GmbH, all rights reserved. &amp;#160; [1] Strozzi, Tazio, et al. "Glacier motion estimation using SAR offset-tracking procedures."&amp;#160;IEEE Transactions on Geoscience and Remote Sensing&amp;#160;40.11 (2002): 2384-2391. &amp;#160; [2] Krainer, Karl, et al. "A 10,300-year-old permafrost core from the active rock glacier Lazaun, southern &amp;#214;tztal Alps (South Tyrol, northern Italy)."&amp;#160;Quaternary Research&amp;#160;83.2 (2015): 324-335.

Abstract. With deepening application of deep learning technology in the field of remote sensing, several challenges persist in the segmentation of remote sensing optical images. These challenges include: (1) insufficient availability of deep learning-based remote sensing semantic segmentation datasets; (2) inadequate utilization of multi-source remote sensing data in the field of semantic segmentation; (3) limited sample size for effective model training, as well as the need to enhance both the speed and accuracy of model training. To address these challenges, this study introduces a multi-source remote sensing dataset consisting of 15000 data pairs, each comprising remote sensing multispectral data, synthetic aperture radar(SAR) images, land use and land cover(LULC) data, digital elevation model (DEM) data as well as the analysis data including Slope, Aspect, and Hillshade. Through the application of an end-to-end network based on pix2pix, effective fusion and feature enhancement of multi-source remote sensing data were achieved. The structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), and Spectral Angle Mapper(SAM) values reached 0.84, 23.14, and 0.19, respectively, representing significant improvements over the baseline Pix2pix model’s performance metrics of 0.62, 18.84, and 0.23. In downstream semantic segmentation applications, the enhanced dataset was utilized to train semantic segmentation models for remote sensing image analysis. This approach effectively improved the training speed and segmentation accuracy of the models, with the mean intersection over union (mIoU) increasing from 0.467 to 0.481 and accuracy rising from 0.734 to 0.746. Moreover, the visual representation of remote sensing image segmentation demonstrated noticeable enhancements.