Fei Jin, Fan Wang, Jie Rui, Zhi Liu, Chao Wang, Hong Zhang
{"title":"Residential area extraction based on conditional generative adversarial networks","authors":"Fei Jin, Fan Wang, Jie Rui, Zhi Liu, Chao Wang, Hong Zhang","doi":"10.1109/BIGSARDATA.2017.8124931","DOIUrl":"https://doi.org/10.1109/BIGSARDATA.2017.8124931","url":null,"abstract":"Automatic extraction of residential area from SAR image is a difficult task due to its complexity. The traditional methods based on segmentation or classification are effective. In this article, we present a novel method based on conditional generative adversarial networks (CGANs) to extract regular residential area in rural and urban region. CGANs is applied to extend the scope of research from supervised learning to semi supervised learning and adversarial training is used to improve the training effect. The experimental results show that the proposed method can achieve better results than traditional methods.","PeriodicalId":334489,"journal":{"name":"2017 SAR in Big Data Era: Models, Methods and Applications (BIGSARDATA)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126123714","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Batch filtering of multi-baseline SAR interferograms","authors":"Chaoying Zhao, Baohang Wang, Qin Zhang, Wu Zhu","doi":"10.1109/BIGSARDATA.2017.8124922","DOIUrl":"https://doi.org/10.1109/BIGSARDATA.2017.8124922","url":null,"abstract":"In current Interferometric synthetic aperture radar (InSAR) applications, small baseline subsets (SBAS) technique is frequently applied by setting both temporal baseline and perpendicular baseline thresholds to select interferograms with less decorrelations. However, in real cases, some high quality interferograms may be discarded and some low quality interferograms may be included based on the former thresholds. On the other hand, with the emergence of Big SAR data era, it is time-consuming and bothersome to determine real high quality interferograms manually or visually. Therefore, this paper proposes a means to determine high quality interferograms based on two thresholds, the one is the coherence of filtered interferograms, and the other is the percentage of high coherence points in interferograms. To this ends, first, robust estimation of coherence matrix of multi-baseline InSAR stack is adopted to restrain some heterogeneous points. Second, coherence matrix for each pixel is decomposed to get dominant phase signal based on the SAR scatterer mechanism. Once the interferograms are determined, phase unwrapping is easily conducted based on the high quality coherent points. Real eight scenes TerraSAR-X data over one land subsidence region is applied to verify of the proposed method.","PeriodicalId":334489,"journal":{"name":"2017 SAR in Big Data Era: Models, Methods and Applications (BIGSARDATA)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128918983","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jiehong Chen, Xi Yan, Lin Zheng, Chao Wang, Hong Zhang
{"title":"The analysis of target parameters in L band full polarimetric data","authors":"Jiehong Chen, Xi Yan, Lin Zheng, Chao Wang, Hong Zhang","doi":"10.1109/BIGSARDATA.2017.8124937","DOIUrl":"https://doi.org/10.1109/BIGSARDATA.2017.8124937","url":null,"abstract":"This paper investigates three target parameters, i.e., target orientation angle, target ellipticity angle, and target phase angle in L band AL O S PAL SAR full polarimetric data collected over Beijing, China. In experiments, the three target parameters are analyzed via histograms, distributed image, and the number of negative power pixels in Freeman decomposition with various compensations. Experimental results show that all the three target parameters have a Gaussian distribution histogram, and conform to the characters of different land covers. After compensation, negative power pixels in Freeman decomposition are dramatically reduced as compared with Freeman decomposition without any target compensations. It is also found that target orientation angle impact on the interaction strongest among the three target parameters, and target ellipticity angle impact on the interaction stronger than target phase angle.","PeriodicalId":334489,"journal":{"name":"2017 SAR in Big Data Era: Models, Methods and Applications (BIGSARDATA)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124719261","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"ALOS-2 PALSAR-2 scansar interferometry for ground deformation monitoring in Beijing-Tianjin area, Northern China","authors":"Yixian Tang, Chao Wang, Hong Zhang","doi":"10.1109/BIGSARDATA.2017.8124919","DOIUrl":"https://doi.org/10.1109/BIGSARDATA.2017.8124919","url":null,"abstract":"Space-borne differential InSAR (DInSAR) technique is an efficient tool for ground deformation monitoring due to its extensive coverage and millimeter level accuracy. With successful operation of Sentienel-1 and ALOS-2, there is increasing interest in exploiting the wide-swath mode of SAR for measuring surface topography and deformation over large area because of its wide coverage and rapid revisit capabilities. In this paper, we present preliminary result of ground deformation analysis in Beijing-Tianjin area by using three ScanSAR mode products of ALOS-2 PALSAR-2 acquired on Sep.10 2015, Oct.22 2015 and Oct.20 2016.","PeriodicalId":334489,"journal":{"name":"2017 SAR in Big Data Era: Models, Methods and Applications (BIGSARDATA)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127325198","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Analysis of oil spill properties based on dual-polarization RADARSAT-2 imagery","authors":"Guannan Li, Ying Li, Yongchao Hou, Tong Wang","doi":"10.1109/BIGSARDATA.2017.8124920","DOIUrl":"https://doi.org/10.1109/BIGSARDATA.2017.8124920","url":null,"abstract":"Marine oil spills is a matter of great concern since it causes environmental disasters and threats the ecological system, thus the capability to detect precisely and analyze sufficiently is a significant guarantee to realize the management followed by oil spill. The Synthetic Aperture Radar (SAR) has recognized as a key remote sensing tool for oil spill monitoring and detecting. This paper uses modified GBVS model to extract oil spill area as the region of interest based on the significant difference between oil-covered region and background sea-water. First, normalize the mean value of the oil spill image, and then filter out the high value by threshold definition; get the final region based the fusion saliency map with GBVS and mean difference. In the end, we discuss four well-known polarization features of oil/sea based on dual-polarization RADARSAT-2 imagery. The study demonstrates that dual-polarization SAR system has real good potential in the study of oil spill detection.","PeriodicalId":334489,"journal":{"name":"2017 SAR in Big Data Era: Models, Methods and Applications (BIGSARDATA)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132135467","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Unsupervised classification of polarimetric SAR images using deep embedding network","authors":"Tianheng Yan, Xudong Sun, Wen Yang, Xinwu Li","doi":"10.1109/BIGSARDATA.2017.8124939","DOIUrl":"https://doi.org/10.1109/BIGSARDATA.2017.8124939","url":null,"abstract":"In this paper we propose an unsupervised polarimetric SAR image classification method using deep embedding network. In this method, we first use superpixel segmen-tation method to produce superpixel regions and use the density peaks clustering (DPC) method to generate the representation points of the superpixel regions. The simi-larity matrix of the representation point and the sample points is then constructed. The low-dimensional manifold features produced by singular value decomposition (SVD) of the similarity matrix are then input onto the deep embedding network which is composed of stacked auto-encoders (SAEs). The unsupervised classification results are finally obtained by clustering algorithm. By using super-pixel segmentation our method introduces spatial constraints. Comparing to other methods, the DPC method guarantees generating more robust representation points. Random mapping of the low-dimensional features in the deep embedding network guarantees the robustness of the method. The experimental results on real polarimetric SAR data demonstrate the effectiveness of the proposed method.","PeriodicalId":334489,"journal":{"name":"2017 SAR in Big Data Era: Models, Methods and Applications (BIGSARDATA)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126984421","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Orbital error modeling and analysis of spaceborne InSAR","authors":"Xue Chen, Junhuan Peng, Honglei Yang","doi":"10.1109/BIGSARDATA.2017.8124930","DOIUrl":"https://doi.org/10.1109/BIGSARDATA.2017.8124930","url":null,"abstract":"Spaceborne synthetic aperture radar interferometry is a key earth observation technique for large-area deformation monitoring. Measurements can be biased by error originating from different sources. Satellite orbital error affects interferograms in the form of an almost linear signal. The commonly used methods to mitigate the orbital error are to calculate the residual baseline error by the Fast Fourier Transform on the residual phase or to fit it with the quadratic surface. These two methods are applicable in most cases, but do not apply in the presence of a large-scale deformation signal that has similar spatial characteristics, such as tectonic movement or tides. Based on the current research on modeling and analysis of baseline error, this paper realized the orbital error network adjustment of Envisat ASAR data in Los Angeles area according to the principle of network adjustment. For quality control on the network level, iterative data snooping has been used, and the result proved its capability to detect and reject outliers.","PeriodicalId":334489,"journal":{"name":"2017 SAR in Big Data Era: Models, Methods and Applications (BIGSARDATA)","volume":" 23","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114051323","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A performance analysis of convolutional neural network models in SAR target recognition","authors":"Jiaqi Shao, C. Qu, Jianwei Li","doi":"10.1109/BIGSARDATA.2017.8124917","DOIUrl":"https://doi.org/10.1109/BIGSARDATA.2017.8124917","url":null,"abstract":"In recent years, the deep learning method represented by Convolutional Neural Network (CNN) has made great progress in the field of image recognition. In this paper, the representative convolution neural network models such as AlexNet, VGGNet, GoogLeNet, ResNet, DenseNet, SENet and so on are applied to SAR image target recognition. According to the accuracy, parameter quantity, training time and other indicators, the performance of different CNN models are analyzed and compared on MSTAR data set, the superiority of CNN model in SAR image target recognition is verified, and it also provides a reference for the follow-up work in this field.","PeriodicalId":334489,"journal":{"name":"2017 SAR in Big Data Era: Models, Methods and Applications (BIGSARDATA)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117180104","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"OpenSARShip 2.0: A large-volume dataset for deeper interpretation of ship targets in Sentinel-1 imagery","authors":"Boying Li, Bin Liu, Lanqing Huang, Weiwei Guo, Zenghui Zhang, Wenxian Yu","doi":"10.1109/BIGSARDATA.2017.8124929","DOIUrl":"https://doi.org/10.1109/BIGSARDATA.2017.8124929","url":null,"abstract":"In the era of big data for synthetic aperture radar (SAR), multiple SAR systems have come into service, including Sentinel-1. The support of large-volume datasets is the key to the deep interpretation of ship targets in Sentinel-1 imagery. In this paper, we present OpenSARShip 2.0, the newest and upgraded version of OpenSARShip, which is dedicated to the deeper interpretation of SAR imagery for marine surveillance. Based on the higher requirements for ship target interpretation, the OpenSARShip 2.0, covering 34528 SAR ship chips with automatic identification system (AIS) information, possesses three improvements: large volume, interference labeling, and type levels, which are described in this paper. In addition, an application case involving the geometric parameter extraction is introduced to demonstrate the applicability of the large dataset.","PeriodicalId":334489,"journal":{"name":"2017 SAR in Big Data Era: Models, Methods and Applications (BIGSARDATA)","volume":"83 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114284419","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Corn mapping uisng multi-temporal fully and compact SAR data","authors":"Lu Xu, Hong Zhang, Chao Wang, Bo Zhang, Meng Liu","doi":"10.1109/BIGSARDATA.2017.8124925","DOIUrl":"https://doi.org/10.1109/BIGSARDATA.2017.8124925","url":null,"abstract":"Corn is one of the most essential economic crops globally and is an important part of China's agricultural industry. Corn growth monitoring is helpful for agricultural land reallocation and production evaluation. The compact polarimetric (CP) SAR could be a good tool for large area surveillance as it overcomes the width limitation of fully-polarimetric (FP) SAR, whereas still inherits abundant information. To testify how CP SAR data behave in corn field mapping, four Radarsat-2 Fine Quad images were continuously acquired to simulate π/4 mode and hybrid mode CP SAR data. Indicators of scattering mechanisms defined by Stokes parameters are analyzed. Comparisons are made between FP and CP SAR data on backscattering coefficients as well. The decision tree is then adopted to extract corn field with temporal information. The results of CP SAR show similar accuracies with FP SAR, indicating good inheritances from FP SAR.","PeriodicalId":334489,"journal":{"name":"2017 SAR in Big Data Era: Models, Methods and Applications (BIGSARDATA)","volume":"129 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127487305","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}