Low-resolution Synthetic Aperture Radar Research Articles

Towards Lightweight Deep Classification for Low-Resolution Synthetic Aperture Radar (SAR) Images: An Empirical Study

Numerous works have explored deep models for the classification of high-resolution natural images. However, limited investigation has been made into a deep classification for low-resolution synthetic aperture radar (SAR) images, which is a challenging yet important task in the field of remote sensing. Existing work adopted ROC–VGG, which has a huge amount of parameters, thus limiting its application in practical deployment. It remains unclear whether the techniques developed in high-resolution natural images to make the model lightweight can be effective for low-resolution SAR images. Therefore, with prior work as the baseline, this work conducts an empirical study, testing three popular lightweight techniques: (1) channel attention module; (2) spatial attention module; (3) multi-stream head. Our empirical results show that these lightweight techniques in the high-resolution natural image domain can also be effective in the low-resolution SAR domain. We reduce the parameters from 9.2M to 0.17M while improving the performance from 94.8% to 96.8%.

Optical-guided residual learning network for synthetic aperture radar image super-resolution

Synthetic aperture radar (SAR) image super-resolution (SR) has broad application prospects in military reconnaissance, marine monitoring, crop monitoring, and other fields. However, when the scale factor is large, current SAR image SR has limitations, such as insufficient feature extraction, difficulty of feature representation, and modeling difficulty of SR reconstruction. These problems have severely restricted the SR performance of SAR images. In addition, optical remote sensing images share higher resolution and richer high-frequency information than SAR remote sensing images. Therefore, we adopt high-resolution (HR) optical images as auxiliary images of SAR image SR to improve the performance of SAR image SR. Our research is conducted from two aspects: the extraction efficiency of HR optical image-guided information and the representation ability of low-resolution (LR) SAR image features. We propose an optical guidance residual networks (OGRNs) based on residual learning for SAR image SR. The OGRN uses optical images during network training, which enhances the performance of SAR image SR, even when optical images are missing during network testing. First, we extract the texture feature of SAR images through the designed residual learning module. Second, we extract the weight map of feature information for SAR image SR from HR optical images by the convolutional layer. Finally, the weight map is used to guide the weighted supervised SR reconstruction for SAR images. Extensive experimental results demonstrated that, under the guidance of optical images, OGRN achieves excellent performance in terms of both quantitative assessment metrics and visual quality.

Airborne SAR Radiometric Calibration Based on Improved Sliding Window Integral Method.

To verify the performance of the high-resolution fully polarimetric synthetic aperture radar (SAR) sensor carried by the Xinzhou 60 remote-sensing aircraft, we used corner reflectors to calibrate the acquired data. The target mechanism in high-resolution SAR images is more complex than it is in low-resolution SAR images, the impact of the point target pointing error on the calibration results is more obvious, and the target echo signal of high-resolution images is more easily affected by speckle noise; thus, more accurate extraction of the point target position and the response energy is required. To solve this problem, this paper introduces image context information and proposes a method to precisely determine the integration region of the corner reflector using sliding windows based on the integral method. The validation indicates that the fully polarimetric SAR sensor on the Xinzhou 60 remote-sensing aircraft can accurately reflect the radiometric characteristics of the ground features and that the integral method can obtain more stable results than the peak method. The sliding window allows the position of the point target to be determined more accurately, and the response energy extracted from the image via the integral method is closer to the theoretical value, which means that the high-resolution SAR system can achieve a higher radiometric calibration accuracy. Additionally, cross-validation reveals that the airborne SAR images have similar quality levels to Sentinel-1A and Gaofen-3 images.

Accurate Segmentation for Low Resolution Satellite images by Discriminative Generative Adversarial Network for Identifying Agriculture Fields

Agricultural field identification is still a difficult issue because of the poor resolution of satellite imagery. Monitoring remote harvest and determining the condition of farmlands rely on the digital approach agricultural applications. Therefore, high-resolution photographs have obtained much more attention since they are more efficient in detecting land cover components. In contrast, because of low-resolution repositories of past satellite images used for time series analysis, wavelet decomposition filter-based analysis, free availability, and economic concerns, low-resolution images are still essential. Using low-resolution Synthetic Aperture Radar (SAR) satellite photos, this study proposes a GAN strategy for locating agricultural regions and determining the crop's cultivation state, linked to the initial or harvesting time. An object detector is used in the preprocessing step of training, followed by a transformation technique for extracting feature information and then the GAN strategy for classifying the crop segmented picture. After testing, the suggested algorithm is applied to the database's SAR images, which are further processed and categorized based on the training results. Using this information, the density between the crops is calculated. After zooming in on SAR photos, the crop condition may be categorized based on crop density and crop distance. The Euclidean distance formula is used to calculate the distance. Finally, the findings are compared to other existing approaches to determine the proposed technique's performance using reliable measures.

Fluvial Features on Titan and Earth: Lessons from Planform Images in Low-resolution SAR

Abstract Cassini Synthetic Aperture Radar (SAR) images of Titan’s surface revealed river networks strikingly similar to those on Earth. However, Cassini SAR has low spatial resolution and image quality compared to data used to map channels on Earth, so traditional methods for characterizing river networks might not yield accurate results on Titan. We mapped terrestrial analog networks with varying resolutions and image qualities to determine which geomorphologic metrics were invariant with scale or resolution. We found that branching angle and drainage density varied significantly with image resolution, and we therefore expect the actual drainage density of Titan’s channel networks to be significantly higher than the values calculated from Cassini data. Calculated network geometry did not change predictably with resolution and would therefore not be an ideal metric for interpreting Titan’s channel networks. The measured channel width, basin length and width, and drainage area all behaved predictably as resolution varied, leading us to conclude that these metrics could be applied to Cassini data. We then mapped all observable fluvial features on Titan—excluding those in the highly incised labyrinth terrains—visible in the Cassini data set. In our new maps, we observe differences in basin shapes between Titan’s polar and equatorial regions and dichotomies in the relative channel density between the northern and southern midlatitudes and poles of Titan: channels are concentrated at the poles and southern midlatitudes. These patterns may reflect differences in bedrock material and/or different climate histories.

Automatic Target Recognition for Synthetic Aperture Radar Images Based on Super-Resolution Generative Adversarial Network and Deep Convolutional Neural Network

Aiming at the problem of the difficulty of high-resolution synthetic aperture radar (SAR) image acquisition and poor feature characterization ability of low-resolution SAR image, this paper proposes a method of an automatic target recognition method for SAR images based on a super-resolution generative adversarial network (SRGAN) and deep convolutional neural network (DCNN). First, the threshold segmentation is utilized to eliminate the SAR image background clutter and speckle noise and accurately extract target area of interest. Second, the low-resolution SAR image is enhanced through SRGAN to improve the visual resolution and the feature characterization ability of target in the SAR image. Third, the automatic classification and recognition for SAR image is realized by using DCNN with good generalization performance. Finally, the open data set, moving and stationary target acquisition and recognition, is utilized and good recognition results are obtained under standard operating condition and extended operating conditions, which verify the effectiveness, robustness, and good generalization performance of the proposed method.

Nonparametric blind SAR image super resolution based on combination of the compressive sensing and sparse priors

This paper by proposing a novel approach, is one the first works that addresses the highly ill-posed problem of nonparametric blind single image super resolution (SISR) of the synthetic aperture radar (SAR) images. Combination of an adaptive compressive sensing (CS) technique and some effective sparse priors, as a powerful regularizer in the both high resolution (HR) image reconstruction and the point spread function (PSF) estimation domains is the fundamental idea of the proposed method. This task is formulated as a new cost function to be minimized with respect to an intermediate reconstructed HR image patch and a nonparametric PSF kernel, according to the alternative minimization (AM) algorithm. To solve the optimization of cost function, a numerical scheme based on the conjugate gradient least squares (CGLS) method is proposed. Experimental results for the both synthetic and realistic low resolution (LR) SAR images demonstrate that the proposed method achieves the state-of-the-art performance.

Registration of multitemporal low-resolution synthetic aperture radar images based on a new similarity measure

Image registration is concerned with the precise overlap of two images. One challenging problem in this area is the registration of low-resolution synthetic aperture radar (SAR) images. In general, extracting feature points from such images is difficult due to the coarse observation and the severe speckle. The use of area similarity for image registration is another important branch to solve the problem. A similarity measure based on a conditional density function (cdf) is proposed. The cdf is specially tailored for SAR images, where the speckle is generally assumed as multiplicative gamma noise with unit mean. Additionally, a two-step procedure is devised for the registration of intro-model SAR images to improve the computational efficiency. First, the two images are roughly aligned considering only the translational difference. Then small blocks from the two images are accurately aligned and the center point of each block is treated as a control point, which is finally used to obtain the precise affine transformation between the two images. Five SAR image datasets are tested in the experiment part, and the results demonstrate the efficiency and accuracy of the proposed method.

SAR Image Categorization With Log Cumulants of the Fractional Fourier Transform Coefficients

With the advent of high-resolution (HR) synthetic aperture radar (SAR) images from satellites like TerraSAR-X and TanDEM-X, interest is now on patch-oriented image categorization in contrast to the pixel-based classification in low-resolution SAR images. SAR image categorization requires the generation of a compact feature descriptor that accurately defines the content of the image patch under consideration. As phase information plays a critical role in SAR images, this paper proposes the use of a chirplet-derived transform, i.e., the fractional Fourier transform (FrFT), for generating a compact feature descriptor for single-look complex (SLC) SAR images. Representing a SAR signal in rotated joint time-frequency planes via the FrFT allows discovering the underlying backscattering phenomenon of the objects on the ground. SAR image projections on different planes of the joint time-frequency space using the FrFT provide a simple statistical response that is easier to analyze. The proposed method has been compared with a multiscale approach, i.e., Gabor filter banks, a second-order-statistics-based method (as gray-level co-occurrence matrices), and a spectral descriptor method. We demonstrate the suitability of the FrFT-based method for image categorization on the basis of backscattering behavior, whereas the Gabor-filter-bank-based method is found mainly suitable for images with a strong texture. This paper demonstrates enhancement in the separability for most of the considered categories when using logarithmic cumulants instead of linear moments for both the FrFT-based and Gabor-filter-bank-based methods. The experimental database consists of 2000 image patches (of size 200 × 200 pixels) extracted from SLC HR TerraSAR-X scenes.

Timely Low Resolution SAR Imagery To Support Floodplain Modelling: a Case Study Review

It is widely recognised that remote sensing can support flood monitoring, modelling and management. In particular, satellites carrying Synthetic Aperture Radar (SAR) sensors are valuable as radar wavelengths can penetrate cloud cover and are insensitive to daylight. However, given the strong inverse relationship between spatial resolution and revisit time, monitoring floods from space in near real time is currently only possible through low resolution (about 100 m pixel size) SAR imagery. For instance, ENVISAT-ASAR (Advanced Synthetic Aperture Radar) in WSM (wide swath mode) revisit times are of the order of 3 days and the data can be obtained within 24 h at no (or low) cost. Hence, this type of space-borne data can be used for monitoring major floods on medium-to-large rivers. This paper aims to discuss the potential for, and uncertainties of, coarse resolution SAR imagery to monitor floods and support hydraulic modelling. The paper first describes the potential of globally and freely available space-borne data to support flood inundation modelling in near real time. Then, the uncertainty of SAR-derived flood extent maps is discussed and the need to move from deterministic binary maps (wet/dry) of flood extent to uncertain flood inundation maps is highlighted.

A wavelet transform coder supporting browsing and transmission of sea ice SAR imagery

A wavelet-transform-based algorithm for sea ice synthetic aperture radar (SAR) image data compression is presented. Compression of the relatively low resolution (100 m) SAR data is necessary to enable the transmission of such images from the Finnish Ice Service to users in the Baltic Sea. On board the ships, the SAR images are used for navigation purposes. Hence, in the visual appearance of the compressed image, special attention must be paid to the sharpness of small details. Several target-dependent features are incorporated in the compression scheme developed. The addition of these features was motivated by an examination of the wavelet coefficient statistics for the SAR data. In the quantization phase, the sensitivity characteristics of the human visual system were taken into account. As a result, the proposed algorithm deviates in several respects from the standard procedures used in wavelet-based compression. The algorithm requires the setting up of multiple user-defined parameters to satisfy the requirements of the users and data. This requires supervision of an expert, but also makes it flexible for many kinds of data and user requirements. The algorithm presented gives satisfactory results with a compression ratio of 20:1. Since only the location of ice-covered areas is of significance for ship traffic, the algorithm introduced contains an option to mask off open sea areas with the aid of an automatic open-water detection procedure. We also report the results of a user evaluation, in which the proposed algorithm is compared to the algorithm currently used by the Finnish Ice Service, as well as to the JPEG standard. The results of the evaluation favor the use of the proposed algorithm.

Target detection in SAR imagery by genetic programming

The automatic detection of ships in low-resolution synthetic aperture radar (SAR) imagery is investigated in this article. The detector design objectives are to maximise detection accuracy across multiple images, to minimise the computational effort during image processing, and to minimise the effort during the design stage. The results of an extensive numerical study show that a novel approach, using genetic programming (GP), successfully evolves detectors which satisfy the earlier objectives. Each detector represents an algebraic formula and thus the principles of detection can be discovered and reused. This is a major advantage over artificial intelligence techniques which use more complicated representations, e.g. neural networks.

Satellite observation and climate system model simulation of the St. Lawrence Island polynya

The St. Lawrence Island polynya (SLIP) is a commonly occurring winter phenomenon in the Bering Sea, in which dense saline water produced during new ice formation is thought to flow northward through the Bering Strait to help maintain the Arctic Ocean halocline. Winter darkness and inclement weather conditions have made continuous in situ and remote observation of this polynya difficult. However, imagery acquired from the European Space Agency ERS-1 Synthetic Aperture Radar (SAR) has allowed observation of the St. Lawrence Island polynya using both the imagery and derived ice displacement products. With the development of ARCSyM, a high resolution regional model of the Arctic atmosphere/sea ice system, simulation of the SLIP in a climate model is now possible. Intercomparisons between remotely sensed products and simulations can lead to additional insight into the SLIP formation process. Low resolution SAR, SSM/I and AVHRR infrared imagery for the St. Lawrence Island region are compared with the results of a model simulation for the period of 24—27 February 1992. The imagery illustrates a polynya event (polynya opening). With the northerly winds strong and consistent over several days, the coupled model captures the SLIP event with moderate accuracy. However, the introduction of a stability dependent atmosphere-ice drag coefficient, which allows feedbacks between atmospheric stability, open water, and air-ice drag, produces a more accurate simulation of the SLIP in comparison to satellite imagery. Model experiments show that the polynya event is forced primarily by changes in atmospheric circulation followed by persistent favorable conditions: ocean surface currents are found to have a small but positive impact on the simulation which is enhanced when wind forcing is weak or variable.