Gaofen-3 Synthetic Aperture Radar Research Articles

Intelligent Wind Retrieval from Chinese Gaofen-3 SAR Imagery in Quad Polarization

AbstractThis study proposes the use of the artificial neural network for wind retrieval with ChineseGaofen-3(GF-3) synthetic aperture radar (SAR) data. More than 10 000 images acquired in wave mode and quad-polarization strip map were collected over global seas throughout the 2-yr mission. TheGF-3operated in a quad-polarization channel—vertical–vertical (VV), vertical–horizontal (VH), horizontal–horizontal (HH), and horizontal–vertical (HV). These images were collocated with winds from the European Centre for Medium-Range Weather Forecasts at a 0.125° grid. The newly released wind retrieval algorithm for copolarization (VV and HH) SAR included CMOD7 and C-SARMOD2. We developed an algorithm based on an artificial neural network method using the SAR-measured normalized radar cross section at quad-polarization channels, herein named QPWIND_GF. Simulations using the QPWIND_GF showed that the correlation coefficient of wind speed was 0.94. We then validated the retrieval wind speeds against the measurements at a 0.25° grid from the Advanced Scatterometer. A comparison showed that the root-mean-square error (RMSE) of wind speed was 0.74 m s−1, which was better than the wind speed obtained using state-of-the-art methods—including, for example, CMOD7 (RMSE 0.88 m s−1) and C-SARMOD2 (RMSE 1.98 m s−1). The finding indicated that the accuracy of wind retrieval fromGF-3SAR images was significantly improved. Our work demonstrates the advanced feasibility of an artificial neural network method for SAR marine applications.

Efficient Parameter Estimation for Sparse SAR Imaging Based on Complex Image and Azimuth-Range Decouple.

Sparse signal processing theory has been applied to synthetic aperture radar (SAR) imaging. In compressive sensing (CS), the sparsity is usually considered as a known parameter. However, it is unknown practically. For many functions of CS, we need to know this parameter. Therefore, the estimation of sparsity is crucial for sparse SAR imaging. The sparsity is determined by the size of regularization parameter. Several methods have been presented for automatically estimating the regularization parameter, and have been applied to sparse SAR imaging. However, these methods are deduced based on an observation matrix, which will entail huge computational and memory costs. In this paper, to enhance the computational efficiency, an efficient adaptive parameter estimation method for sparse SAR imaging is proposed. The complex image-based sparse SAR imaging method only considers the threshold operation of the complex image, which can reduce the computational costs significantly. By utilizing this feature, the parameter is pre-estimated based on a complex image. In order to estimate the sparsity accurately, adaptive parameter estimation is then processed in the raw data domain, combining with the pre-estimated parameter and azimuth-range decouple operators. The proposed method can reduce the computational complexity from a quadratic square order to a linear logarithm order, which can be used in the large-scale scene. Simulated and Gaofen-3 SAR data processing results demonstrate the validity of the proposed method.

Semi-Empirical Algorithm for Wind Speed Retrieval from Gaofen-3 Quad-Polarization Strip Mode SAR Data

Synthetic aperture radar (SAR) is a suitable tool to obtain reliable wind retrievals with high spatial resolution. The geophysical model function (GMF), which is widely employed for wind speed retrieval from SAR data, describes the relationship between the SAR normalized radar cross-section (NRCS) at the copolarization channel (vertical-vertical and horizontal-horizontal) and a wind vector. SAR-measured NRCS at cross-polarization channels (horizontal-vertical and vertical-horizontal) correlates with wind speed. In this study, a semi-empirical algorithm is presented to retrieve wind speed from the noisy Chinese Gaofen-3 (GF-3) SAR data with noise-equivalent sigma zero correction using an empirical function. GF-3 SAR can acquire data in a quad-polarization strip mode, which includes cross-polarization channels. The semi-empirical algorithm is tuned using acquisitions collocated with winds from the European Center for Medium-Range Weather Forecasts. In particular, the proposed algorithm includes the dependences of wind speed and incidence angle on cross-polarized NRCS. The accuracy of SAR-derived wind speed is around 2.10 m s-1 root mean square error, which is validated against measurements from the Advanced Scatterometer onboard the Metop-A/B and the buoys from the National Data Buoy Center of the National Oceanic and Atmospheric Administration. The results obtained by the proposed algorithm considering the incidence angle in a GMF are relatively more accurate than those achieved by other algorithms. This work provides an alternative method to generate operational wind products for GF-3 SAR without relying on ancillary data for wind direction.

Analysis of polarisation features of typical targets in Chengdu using GF‐3 fully polarimetric synthetic aperture radar data

The fully polarimetric synthetic aperture radar (SAR) data contain much backscatter intensity information and have great advantage for land monitoring. Analysing the polarisation features of typical targets is a critical step for landing monitoring, especially the ground objects identification, and thus accurately obtaining and analysing the polarisation features is of importance. Gaofen-3(GF-3) is the first C band fully polarimetric SAR satellite in China, which is widely used in various fields such as land monitoring, disaster reduction and so on. Here, the GF-3 fully polarimetric SAR data at three time-slices were used to obtain and analyse the polarisation features of typical targets in Chengdu. The results revealed the scattering type of the typical targets including artificial buildings and vegetation and indicated that the GF-3 SAR data were suitable to be used for obtaining polarisation features of typical targets and have the potential for discriminate artificial building and further for landing monitoring.

Assessments of Ocean Wind Retrieval Schemes Used for Chinese Gaofen-3 Synthetic Aperture Radar Co-Polarized Data

This paper assesses different retrieval schemes used for the Chinese Gaofen-3 Synthetic Aperture Radar (GF-3 SAR) co-polarized data. The data consist of 4186 GF-3 data points and collocated wind information from sources including the ASCAT scatterometer, HY2A-SCAT scatterometer, and National Data Buoy Center (NDBC) buoy wind data set. The VV-polarized geophysical model function (GMF) is a CMOD7 model while the HH-polarized GMF is a hybrid of the CMOD7 and PR model. Assessments involve comparisons between SAR-derived and collocated winds in terms of the root-mean-square difference (RMSD) and bias. First, a comparison between the two retrieval schemes for the VV-polarized data clearly shows that the optimal scheme performs better than the classical scheme for wind speed retrieval. Comparisons for HH-polarized data show similar results. These experiments indicate that the wind speed RMSDs for the GF-3 co-polarized data are within 2 m/s when using the optimal scheme. Moreover, the wind direction RMSDs from the two schemes have no significant difference, with values near 20°. Overall, these assessments indicate that the GF-3 co-polarized data are sufficient for operational wind speed retrieval using the optimal scheme. However, wind direction retrieval requires further improvement.

Random cross-observation intensity consistency method for large-scale SAR images mosaics: An example of Gaofen-3 SAR images covering China

Large-scale microwave remote sensing research requires multiple synthetic aperture radar (SAR) image mosaics. However, owing to multiple factors, different SAR images frequently exhibit intensity differences. This leads to poor intensity continuity after mosaicking, making SAR image interpretation extremely difficult. The existing method assumes that SAR images have been subject to strict radiometric calibration and mainly focuses on intensity balancing to minimize the intensity difference between adjacent images. This method is not effective when the intensity difference between SAR images is large. In this paper, a novel random cross-observation intensity consistency (RCOIC) algorithm is proposed to eliminate the intensity differences for SAR image mosaics. The proposed algorithm is based on global intensity constraints. First, source images are downsampled. Then, the gain correction coefficients of each image in a study area are accurately estimated through random cross-observation. Further, multi-view local intensity correction of each image in the study area is performed using the overlapping areas of adjacent images, and a low-resolution intensity reference map covering the study area is generated. Finally, the intensities of high-resolution source images are corrected by constructing local independent linear models of images based on the low-resolution intensity reference map. Gaofen-3 satellite images with large ranges and large time differences, covering most of China's land area, were selected to test the effect of the newly developed algorithm. After processing, the overall visual effect was good after mosaicking, the intensity transitions at the edges of the images were smooth, the intensity distributions corresponded better to actual ground objects, and the statistical result of adjacent images overlapping regions was good. This validates the reliability of the algorithm.

Design and Implementation of a Multi-Band Active Radar Calibrator for SAR

Over the past decade, IECAS (Institute of Electronics, Chinese Academy of Sciences) has developed a set of L-, S-, C-, and X-band active radar calibrators that are deployed during the calibration campaigns for HJ1C synthetic aperture radar (SAR), Gaofen-3 SAR, and so on. In the near future, P-band and Ka-band spaceborne SARs will be launched. We found that it is not convenient to develop special active radar calibrators (ARCs) for a specific SAR or a specific frequency band SAR, and the acquired experience could help in the design and development of a multi-band ARC. This paper describes the design and implementation of a multi-band active radar calibrator which can operate in the L-, C-, X-, and Ka-bands. Moreover, laboratory measurements are performed to characterize the performance of the multi-band ARC, paying particular attention to the gain stability, the system transfer function, the gain flatness, and the linearity of the ARC receiver. Three such ARCs are developed, and to our knowledge, the multi-band ARC is the first of its kind in China or even in the world, and it can be used to implement the calibration campaigns of the Chinese Gaofen-3 SAR, Shenzhen-1 SAR, Luojia-2 SAR, and so on.

Calibration of the Copolarized Backscattering Measurements From Gaofen-3 Synthetic Aperture Radar Wave Mode Imagery

Accurate absolute radiometric calibration of spaceborne synthetic aperture radar (SAR) sensor is of great importance in quantitative oceanic monitoring. Traditionally, the calibration constant is determined by analyzing measurements of man-made calibrators deployed on land at a high expense. In this paper, a technique called numerical weather prediction (NWP)-based ocean calibration method for estimating the calibration constant on the basis of SAR observations over ocean is applied to the first Chinese C-band SAR satellite Gaofen-3 (GF-3). The ocean calibration is performed on GF-3 wave mode SAR images at VV and HH polarizations over a period of one year from September 2017 to August 2018. Verification against the independent scatterometer-derived winds shows that after NWP-based ocean recalibration, the accuracy of wind speed retrieval from GF-3 wave mode imagery could reach the wind speed estimation performance using the state-of-the-art of SAR data. These results indicate that GF-3 wind retrievals are promising for operational oceanic products and applications if the SAR data are appropriately calibrated by the proposed NWP-based ocean calibration approach.

Demonstration of Time-Series InSAR Processing in Beijing Using a Small Stack of Gaofen-3 Differential Interferograms

More and more synthetic aperture radar (SAR) satellites in orbit provide abundant data for remote sensing applications. In August 2016, China launched a new Earth observation SAR satellite, Gaofen-3 (GF-3). In this paper, we utilize a small stack of GF-3 differential interferograms to map land subsidence in Beijing (China) using the time-series SAR interferometry (InSAR) technique. The small stack of differential interferograms is generated with 5 GF-3 SAR images from March 2017 to January 2018. Orbit errors are carefully addressed and removed during differential InSAR (DInSAR) processing. Truncated singular-value decomposition (TSVD) is applied to strengthen the robustness of deformation rate estimation. To validate the results of GF-3 data, an additional deformation measurement using 26 Sentinel-1B images from March 2017 to February 2018 is carried out using the persistent scatterer interferometry (PSI) technique. By implementing a cross-comparison, we find that the retrieved results from GF-3 images and Sentinel-1 images are spatially consistent. The standard deviation of vertical deformation rate differences between two data stacks is 11.24 mm/y in the study area. The results shown in this paper demonstrate the reasonable potential of GF-3 SAR images to monitor land subsidence.

Intertidal area classification with generalized extreme value distribution and Markov random field in quad-polarimetric synthetic aperture radar imagery

Classification of intertidal area in synthetic aperture radar (SAR) images is an important yet challenging issue when considering the complicatedly and dramatically changing features of tidal fluctuation. The difficulty of intertidal area classification is compounded because a high proportion of this area is frequently flooded by water, making statistical modeling methods with spatial contextual information often ineffective. Because polarimetric entropy and anisotropy play significant roles in characterizing intertidal areas, in this paper we propose a novel unsupervised contextual classification algorithm. The key point of the method is to combine the generalized extreme value (GEV) statistical model of the polarization features and the Markov random field (MRF) for contextual smoothing. A goodness-of-fit test is added to determine the significance of the components of the statistical model. The final classification results are obtained by effectively combining the results of polarimetric entropy and anisotropy. Experimental results of the polarimetric data obtained by the Chinese Gaofen-3 SAR satellite demonstrate the feasibility and superiority of the proposed classification algorithm.

Ship Classification and Detection Based on CNN Using GF-3 SAR Images

Ocean surveillance via high-resolution Synthetic Aperture Radar (SAR) imageries has been a hot issue because SAR is able to work in all-day and all-weather conditions. The launch of Chinese Gaofen-3 (GF-3) satellite has provided a large number of SAR imageries, making it possible to marine targets monitoring. However, it is difficult for traditional methods to extract effective features to classify and detect different types of marine targets in SAR images. This paper proposes a convolutional neutral network (CNN) model for marine target classification at patch level and an overall scheme for marine target detection in large-scale SAR images. First, eight types of marine targets in GF-3 SAR images are labelled based on feature analysis, building the datasets for further experiments. As for the classification task at patch level, a novel CNN model with six convolutional layers, three pooling layers, and two fully connected layers has been designed. With respect to the detection part, a Single Shot Multi-box Detector with a multi-resolution input (MR-SSD) is developed, which can extract more features at different resolution versions. In order to detect different targets in large-scale SAR images, a whole workflow including sea-land segmentation, cropping with overlapping, detection with MR-SSD model, coordinates mapping, and predicted boxes consolidation is developed. Experiments based on the GF-3 dataset demonstrate the merits of the proposed methods for marine target classification and detection.

Capabilities of Chinese Gaofen-3 Synthetic Aperture Radar in Selected Topics for Coastal and Ocean Observations

Gaofen-3 (GF-3), the first Chinese spaceborne synthetic aperture radar (SAR) in C-band for civil applications, was launched on August 2016. Some studies have examined the use of GF-3 SAR data for ocean and coastal observations, but these studies generally focus on one particular application. As GF-3 has been in operation over two years, it is essential to evaluate its performance in ocean observation, a primary goal of the GF-3 launch. In this paper, we offer an overview demonstrating the capabilities of GF-3 SAR in ocean and coastal observations by presenting several representative cases, i.e., the monitoring of intertidal flats, offshore tidal turbulent wakes and oceanic internal waves, to highlight the GF-3’s full polarimetry, high spatial resolution and wide-swath imaging advantages. Moreover, we also present a detailed analysis of the use of GF-3 quad-polarization data for sea surface wind retrievals and wave mode data for sea surface wave retrievals. The case studies and statistical analysis suggest that GF-3 has good ocean and coastal monitoring capabilities, though further improvements are possible, particularly in radiometric calibration and stable image quality.

Block Adjustment without GCPs for Chinese Spaceborne SAR GF-3 Imagery.

The Gaofen-3 (GF-3) satellite is the first C-band multi-polarization synthetic aperture radar (SAR) with the ability of high-accuracy mapping in China. However, the Ground Control Points (GCPs) are essential to ensure the accuracy of mapping for GF-3 SAR imagery at present. In this paper, we analyze the error sources that affect the geometric processing and propose a new block adjustment method without GCPs for GF-3 SAR imagery. Firstly, the geometric calibration of GF-3 image is carried out. Secondly, the rational polynomial coefficient (RPC) model is directly generated after the geometric calibration parameters compensation of each image. Finally, we solve the orientation parameters of the GF-3 images through DEM assisted planar block adjustment and conduct ortho-rectification. With two different imaging modes of GF-3 satellite, which include the QPSI and FS2, we carry out the block adjustment without GCPs. Experimental results of testing areas including Wuhan city and Hubei province in China show that the geometric mosaic accuracy and the absolute positioning accuracy of the orthophoto are better than one pixel, which has laid a good foundation for the application of GF-3 image in global high-accuracy mapping.

Evaluation of Chinese Quad-polarization Gaofen-3 SAR Wave Mode Data for Significant Wave Height Retrieval

Our work describes the accuracy of Chinese quad-polarization Gaofen-3 (GF-3) synthetic aperture radar (SAR) wave mode data for wave retrieval and provides guidance for the operational applications of GF-3 SAR. In this study, we evaluated the accuracy of the SAR-derived significant wave height (SWH) from 10,514 GF-3 SAR images with visible wave streaks acquired in wave mode by using the existing wave retrieval algorithms, e.g., the theoretical-based algorithm parameterized first-guess spectrum method (PFSM), the empirical algorithm CSAR_WAVE2 for VV-polarization, and the algorithm for quad-polarization (Q-P). The retrieved SWHs were compared with the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis field with 0.125° grids. The root mean square error (RMSE) of the SWH is 0.57 m, found using CSAR_WAVE2, and this RMSE value was less than the RMSE values for the analysis results achieved with the PFSM and Q-P algorithms. The statistical analysis also indicated that wind speed had little impact on the bias with increasing wind speed. However, the retrieval tended to overestimate when the SWH was smaller than 2.5 m and underestimate with an increasing SWH. This behavior provides a perspective of the improvement needed for the SWH retrieval algorithm using the GF-3 SAR acquired in wave mode.

Evaluation of Wind Retrieval from Co-Polarization Gaofen-3 SAR Imagery Around China Seas

Gaofen-3 (GF-3) is the first Chinese space-borne satellite to carry the C-band multi-polarization synthetic aperture radar (SAR). Marine applications, i.e., winds and waves retrieved from GF-3 SAR images, have been operational since January 2017. In this study, we have collected more than 1000 quad-polarization (vertical-vertical (VV); horizontal-horizontal (HH); vertical- horizontal (VH); horizontal-vertical (HV)) GF-3 SAR images, which were acquired around the China Seas from September 2016 to September 2017. Wind streaks were visible in these images in co-polarization (VV and HH) channel. Geophysical model functions (GMFs), including the CMOD5N together with polarization ratio (PR) model and C-SARMOD, were used to retrieve winds from the collected co-polarization GF-3 SAR images. Wind directions were directly obtained from GF-3 SAR images. Then, the SAR-derived wind speeds were compared with the measurements at a 0.25° grid from the Advanced Scatterometer on board the Metop-A/B and microwave radiometer WindSAT. Based on the analysis, empirical corrections are proposed to improve the performance of the two GMFs. Results of this study show that the standard deviation of wind speed is 1.63 m s−1 with a 0.19 m s−1 bias and 1.71 m s−1 with a 0.26 m s−1 bias for VV- and HH-polarization GF-3 SAR, respectively. Our work not only systematically evaluates wind retrieval by using the two advanced GMFs and PR models but also proposes empirical corrections to improve the accuracy of wind retrievals from GF-3 SAR images around the China Seas and thus enhance the accuracy of near real-time operational SAR-derived wind products.

A Promising Method of Typhoon Wave Retrieval from Gaofen-3 Synthetic Aperture Radar Image in VV-Polarization.

The motivation of this work is to explore the possibility of typhoon wave retrieval (the main parameter is significant wave height (SWH)) for C-band Gaofen (GF-3) synthetic aperture radar (SAR) with a wide swath coverage (>400 km). We aim to establish an analysis of a typhoon wave in the subresolution-scale (approximately 20 × 20 km2) on GF-3 SAR through SAR-measured parameters, including a normalized radar cross section (NRCS) and variance of the normalized SAR image (herein called cvar), which are the basic variables in an empirical wave retrieval algorithm and are independent of visible wave streaks. Several typhoons around the China Seas were captured by Chinese GF-3 SAR in 2017; e.g., Noru, Doksuri, Talim and Hato. The wave fields simulated from the third-generation numerical wave model WAVEWATCH-III (WW3) are collocated with these images. In general, the distribution patterns of the typhoon waves from the WW3 model are consistent with wave fields from the European Centre for Medium-Range Weather Forecasts (ECMWF) at 0.125° grids, indicating that the simulation results from the WW3 model are suitable for our study. In addition to winds retrieved from GF-3 SAR images in vertical-horizontal (VH) polarization, the characteristics of the typhoon wave on vertical-vertical (VV) polarization GF-3 SAR images are studied. It is found that SWH has a linear relationship with NRCS and cvar, however, SWH fluctuates with wind speed at all incidence angles. Based on the analyzed results, we simply tune two empirical wave retrieval algorithms for GF-3 SAR in typhoons. Although the correlation (COR) reaches 0.5 taking account into the NRCS term, a more accurate retrieval algorithm, including more related terms, is anticipated for further development for GF-3 SAR and validated through more typhoon images.

Validation of significant wave height retrieval from co-polarization Chinese Gaofen-3 SAR imagery using an improved algorithm

Chinese Gaofen-3 (GF-3) is the first civilian satellite to carry C-band (5.3 GHz) synthetic aperture radar (SAR). During the period of August 2016 to December 2017, 1 523 GF-3 SAR images acquired in quad-polarization (vertical-vertical (VV), horizontal-horizontal (HH), vertical-horizontal (VH), and horizontal-vertical (HV)) mode were recorded, mostly around China’s seas. In our previous study, the root mean square error (RMSE) of significant wave height (SWH) was found to be around 0.58 m when compared with retrieval results from a few GF-3 SAR images in co-polarization (VV and HH) with moored measurements by using an empirical algorithm CSAR_WAVE. We collected a number of sub-scenes from these 1 523 images in the co-polarization channel, which were collocated with wind and SWH data from the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis field at a 0.125° grid. Through the collected dataset, an improved empirical wave retrieval algorithm for GF-3 SAR in co-polarization was tuned, herein denoted as CSAR_WAVE2. An additional 92 GF-3 SAR images were implemented in order to validate CSAR_WAVE2 against SWH from altimeter Jason-2, showing an about 0.52 m RMSE of SWH for co-polarization GF-3 SAR. Therefore, we conclude that the proposed empirical algorithm has a good performance for wave retrieval from GF-3 SAR images in co-polarization.

Water Detection in Urban Areas from GF-3.

The rapid and accurate detection of urban water is critical for urban management, river detection, and flood disaster assessment. This study is devoted to detecting water by GaoFen-3 (GF-3) Synthetic Aperture Radar (SAR) images with high spatial resolution. There have been no effective solutions that discriminate water and building shadows using a single SAR image in previous research. Inspired by the principle that every shadow has a corresponding building nearby, a new method is proposed in this study, whereby building shadows are removed depending on the correspondence of buildings and their shadows. The proposed method is demonstrated effective and efficient by experimental results on six GF-3 SAR images. The Receiver Operating Characteristic (ROC) curves of the water detection results indicate that the proposed method increases the Probability of Detection (PD) to 98.36% and decreases the Probability of False Alarm (PFA) to 1.91% compared with the thresholding method, where, at the same PFA level, the maximum PD of the thresholding method is 72.62% in all testing samples. The proposed method is capable of removing building shadows and detecting water with high precision in urban areas, which presents the great potential of high-spatial-resolution GF-3 images in terms of water resource management.

Development of Wind Speed Retrieval from Cross-Polarization Chinese Gaofen-3 Synthetic Aperture Radar in Typhoons.

The purpose of our work is to determine the feasibility and effectiveness of retrieving sea surface wind speeds from C-band cross-polarization (herein vertical-horizontal, VH) Chinese Gaofen-3 (GF-3) SAR images in typhoons. In this study, we have collected three GF-3 SAR images acquired in Global Observation (GLO) and Wide ScanSAR (WSC) mode during the summer of 2017 from the China Sea, which includes the typhoons Noru, Doksuri and Talim. These images were collocated with wind simulations at 0.12° grids from a numeric model, called the Regional Assimilation and Prediction System-Typhoon model (GRAPES-TYM). Recent research shows that GRAPES-TYM has a good performance for typhoon simulation in the China Sea. Based on the dataset, the dependence of wind speed and of radar incidence angle on normalized radar cross (NRCS) of VH-polarization GF-3 SAR have been investigated, after which an empirical algorithm for wind speed retrieval from VH-polarization GF-3 SAR was tuned. An additional four VH-polarization GF-3 SAR images in three typhoons, Noru, Hato and Talim, were investigated in order to validate the proposed algorithm. SAR-derived winds were compared with measurements from Windsat winds at 0.25° grids with wind speeds up to 40 m/s, showing a 5.5 m/s root mean square error (RMSE) of wind speed and an improved RMSE of 5.1 m/s wind speed was achieved compared with the retrieval results validated against GRAPES-TYM winds. It is concluded that the proposed algorithm is a promising potential technique for strong wind retrieval from cross-polarization GF-3 SAR images without encountering a signal saturation problem.

Ship Detection in Gaofen-3 SAR Images Based on Sea Clutter Distribution Analysis and Deep Convolutional Neural Network.

Target detection is one of the important applications in the field of remote sensing. The Gaofen-3 (GF-3) Synthetic Aperture Radar (SAR) satellite launched by China is a powerful tool for maritime monitoring. This work aims at detecting ships in GF-3 SAR images using a new land masking strategy, the appropriate model for sea clutter and a neural network as the discrimination scheme. Firstly, the fully convolutional network (FCN) is applied to separate the sea from the land. Then, by analyzing the sea clutter distribution in GF-3 SAR images, we choose the probability distribution model of Constant False Alarm Rate (CFAR) detector from K-distribution, Gamma distribution and Rayleigh distribution based on a tradeoff between the sea clutter modeling accuracy and the computational complexity. Furthermore, in order to better implement CFAR detection, we also use truncated statistic (TS) as a preprocessing scheme and iterative censoring scheme (ICS) for boosting the performance of detector. Finally, we employ a neural network to re-examine the results as the discrimination stage. Experiment results on three GF-3 SAR images verify the effectiveness and efficiency of this approach.