Pub Date : 2017-05-01DOI: 10.1109/RSIP.2017.7958805
Chunsen Zhang, Guojun Li, W. Cui
The change detection method based on multi-temporal object was implemented by chi-square test and Gaussian distribution iteration to find the changed object in the past. However, trapped in the sample data does not obey the Gaussian distribution, the detection effect is not ideal. In order to fix this problem, a method based on law of cosines with box-whisker plot is proposed. First, the feature space of different time images is constructed. Then, the law of cosines is used to calculate the change index of every object. The changed objects are identified through analyzing the change index by the box-whisker plot at last. High-resolution remote sensing images of GF-1 are used as the experimental data. The experimental results show that the correct detection accuracy and omissions rate accuracy are much better than the results of the traditional multi-temporal object based change detection.
{"title":"High resolution remote sensing image change detection based on law of cosines with box-whisker plot","authors":"Chunsen Zhang, Guojun Li, W. Cui","doi":"10.1109/RSIP.2017.7958805","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958805","url":null,"abstract":"The change detection method based on multi-temporal object was implemented by chi-square test and Gaussian distribution iteration to find the changed object in the past. However, trapped in the sample data does not obey the Gaussian distribution, the detection effect is not ideal. In order to fix this problem, a method based on law of cosines with box-whisker plot is proposed. First, the feature space of different time images is constructed. Then, the law of cosines is used to calculate the change index of every object. The changed objects are identified through analyzing the change index by the box-whisker plot at last. High-resolution remote sensing images of GF-1 are used as the experimental data. The experimental results show that the correct detection accuracy and omissions rate accuracy are much better than the results of the traditional multi-temporal object based change detection.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"2000 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129542147","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}
Pub Date : 2017-05-01DOI: 10.1109/RSIP.2017.7958816
Zhao Chen, Muhammad Sohail, Bin Wang
Change detection (CD) for multitemporal hyperspectral images (HSI) consists of two steps, change feature extraction and identification. This paper proposes a novel spectrally-spatially regularized low-rank and sparse decomposition model (LRSD_SS), to extract clean change features from corrupted spectral change vectors (SCV) of multitemporal HSI. It decomposes SCV into spatially smoothed low-rank data, sparse outliers and Gaussian noise. The experimental results validate the effectiveness and the efficiency of LRSD_SS.
{"title":"Low-rank matrix decomposition with a spectral-spatial regularization for change detection in hyperspectral imagery","authors":"Zhao Chen, Muhammad Sohail, Bin Wang","doi":"10.1109/RSIP.2017.7958816","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958816","url":null,"abstract":"Change detection (CD) for multitemporal hyperspectral images (HSI) consists of two steps, change feature extraction and identification. This paper proposes a novel spectrally-spatially regularized low-rank and sparse decomposition model (LRSD_SS), to extract clean change features from corrupted spectral change vectors (SCV) of multitemporal HSI. It decomposes SCV into spatially smoothed low-rank data, sparse outliers and Gaussian noise. The experimental results validate the effectiveness and the efficiency of LRSD_SS.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121241021","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}
Pub Date : 2017-05-01DOI: 10.1109/RSIP.2017.7958812
Xin Luo, Huan Xie, X. Tong, Haiyan Pan
Water bodies are a fundamental element of urban ecosystems, and water mapping is critical for urban and landscape planning and management. Remote sensing has increasingly been used for water mapping in rural areas; especially, hyperspectral remote sensing image characterized with rich spectrum information provide greater potential for high-accuracy land cover classiflcation, however, the hundreds of bands contained in the image also poses a huge burden on data processing. In this study, aims for water extraction in the densely built urban area, we proposed a fast water extraction method based on spectral analysis of the hyperspectral images. The performance of the new method performs well especially for the extraction of water surface which casts many building shadows. In comparison with the normalized difference water index (NDWI) and K-means classifier, new method obtains significantly higher accuracy than that of NDWI and K-means. Therefore, new method can be used for extracting water with high accuracy, especially in urban areas where shadow caused by high buildings is an important source of classification error.
{"title":"A water extraction method based on airborne hyperspectral images in highly complex urban area","authors":"Xin Luo, Huan Xie, X. Tong, Haiyan Pan","doi":"10.1109/RSIP.2017.7958812","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958812","url":null,"abstract":"Water bodies are a fundamental element of urban ecosystems, and water mapping is critical for urban and landscape planning and management. Remote sensing has increasingly been used for water mapping in rural areas; especially, hyperspectral remote sensing image characterized with rich spectrum information provide greater potential for high-accuracy land cover classiflcation, however, the hundreds of bands contained in the image also poses a huge burden on data processing. In this study, aims for water extraction in the densely built urban area, we proposed a fast water extraction method based on spectral analysis of the hyperspectral images. The performance of the new method performs well especially for the extraction of water surface which casts many building shadows. In comparison with the normalized difference water index (NDWI) and K-means classifier, new method obtains significantly higher accuracy than that of NDWI and K-means. Therefore, new method can be used for extracting water with high accuracy, especially in urban areas where shadow caused by high buildings is an important source of classification error.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133736984","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}
Pub Date : 2017-05-01DOI: 10.1109/RSIP.2017.7958818
Siwei Chen, Chensong Tao
Multi-temporal PolSAR data is suitable for crops classification and growth monitoring. It is still difficult to establish a classifier with good robustness and high generation over a long temporal acquisition duration. This work aims to provide a solution to this task by exploring benefits from both the target scattering mechanism interpretation and the advanced deep learning. A polarimetric-feature-driven deep convolutional neural network classification scheme is established. Comparison studies with multi-temporal UAVSAR datasets validate the efficiency and superiority of the proposal.
{"title":"Multi-temporal PolSAR crops classification using polarimetric-feature-driven deep convolutional neural network","authors":"Siwei Chen, Chensong Tao","doi":"10.1109/RSIP.2017.7958818","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958818","url":null,"abstract":"Multi-temporal PolSAR data is suitable for crops classification and growth monitoring. It is still difficult to establish a classifier with good robustness and high generation over a long temporal acquisition duration. This work aims to provide a solution to this task by exploring benefits from both the target scattering mechanism interpretation and the advanced deep learning. A polarimetric-feature-driven deep convolutional neural network classification scheme is established. Comparison studies with multi-temporal UAVSAR datasets validate the efficiency and superiority of the proposal.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"107 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133061537","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}
Pub Date : 2017-05-01DOI: 10.1109/RSIP.2017.7958799
Yuanyuan Wang, Chao Wang, Hong Zhang
Classification in remote sensing, similar to semantic segmentation in computer vision, is aimed to assign a label to each pixel in images to indicate which class it belongs to. Fully convolutional networks (FCN), one of semantic segmentation methods, is proposed to tackle this problem in fully PolSAR images in this paper. To exploit the polarimetric information in PolSAR images, H-A-α polarimetric decomposition is integrated with FCN. PolSAR images acquired by Gaofen-3, China's SAR satellite, in the C-band with a spatial resolution of 1 meter are utilized. Three variations of FCN, i.e., FCN-32s, FCN-16s, and FCN-8s, and SVM are trained and validated. Experimental results reveal that the both user and product accuracy of the three FCN architectures is more than 2% higher than support vector machine (SVM) for water pixels, 16% higher for vegetation, and 24% higher for the building study areas in the whole image. Besides, the three architectures of FCN are 75 times faster than SVM.
{"title":"Integrating H-A-α with fully convolutional networks for fully PolSAR classification","authors":"Yuanyuan Wang, Chao Wang, Hong Zhang","doi":"10.1109/RSIP.2017.7958799","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958799","url":null,"abstract":"Classification in remote sensing, similar to semantic segmentation in computer vision, is aimed to assign a label to each pixel in images to indicate which class it belongs to. Fully convolutional networks (FCN), one of semantic segmentation methods, is proposed to tackle this problem in fully PolSAR images in this paper. To exploit the polarimetric information in PolSAR images, H-A-α polarimetric decomposition is integrated with FCN. PolSAR images acquired by Gaofen-3, China's SAR satellite, in the C-band with a spatial resolution of 1 meter are utilized. Three variations of FCN, i.e., FCN-32s, FCN-16s, and FCN-8s, and SVM are trained and validated. Experimental results reveal that the both user and product accuracy of the three FCN architectures is more than 2% higher than support vector machine (SVM) for water pixels, 16% higher for vegetation, and 24% higher for the building study areas in the whole image. Besides, the three architectures of FCN are 75 times faster than SVM.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114453830","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}
In this paper, a novel effective method based on Scale-invariant feature transform in Imagery to realize Target tracking, where the discriminating process is improved through Image Matching Processing. It is the first time that the problem of tracking in Imaging processing, contrasted with traditional methods in data processing. It can track target for clutter. Simulation results show that the proposed method has advantages in the efficiency and accuracy under the circumstances with heavy clutter and large measurement error.
{"title":"A novel target tracking method based on scale-invariant feature transform in imagery","authors":"Huang Qinglong, Yun Zhang, Ling Hongbo, Wangbin, Feng Tianjiao","doi":"10.1109/RSIP.2017.7958801","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958801","url":null,"abstract":"In this paper, a novel effective method based on Scale-invariant feature transform in Imagery to realize Target tracking, where the discriminating process is improved through Image Matching Processing. It is the first time that the problem of tracking in Imaging processing, contrasted with traditional methods in data processing. It can track target for clutter. Simulation results show that the proposed method has advantages in the efficiency and accuracy under the circumstances with heavy clutter and large measurement error.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125041670","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}
Pub Date : 2017-05-01DOI: 10.1109/RSIP.2017.7970774
Junchuan Yu, Bokun Yan
As we have entered an era of information, the RS data are undergoing a plosive growth. The needs of large-scale earth observation have led to the development of high-resolution and high-dimensionality RS data, which has posed significant challenges in processing and application. In this paper, we demonstrate some possible solution of large-scale domestic hyperspectral data processing and geological application, mainly from three aspects.
{"title":"Efficient solution of large-scale domestic hyperspectral data processing and geological application","authors":"Junchuan Yu, Bokun Yan","doi":"10.1109/RSIP.2017.7970774","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7970774","url":null,"abstract":"As we have entered an era of information, the RS data are undergoing a plosive growth. The needs of large-scale earth observation have led to the development of high-resolution and high-dimensionality RS data, which has posed significant challenges in processing and application. In this paper, we demonstrate some possible solution of large-scale domestic hyperspectral data processing and geological application, mainly from three aspects.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"2005 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128230829","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}
Pub Date : 2017-05-01DOI: 10.1109/RSIP.2017.7958814
D. Guo, Bo Chen
In this letter, a novel deep-leaming-based feature selection method based on Poisson Gamma Belief Network (PGBN), is proposed to extract multi-layer feature from SAR images data. As a deep Bayesian generative network, PGBN has the ability to extract a multilayer structured representation from the complex SAR images owing to the existence of Poisson likelihood and multilayer gamma hidden variables, at the same time the PGBN can be viewed as a deep non-negative matrix factorization model. Note that the PGBN model is an unsupervised deep generative network and it fails to make full use of the label information in training stage. Therefore, the NBPGBN model is further proposed to obtain a higher recognition performance and training efficiency based on Naïve Bayes rule. The experimental results on MSTAR dataset show that the feature extracted by this new approach has better structured information and perform better classification result compared with some related algorithms.
{"title":"SAR image target recognition via deep Bayesian generative network","authors":"D. Guo, Bo Chen","doi":"10.1109/RSIP.2017.7958814","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958814","url":null,"abstract":"In this letter, a novel deep-leaming-based feature selection method based on Poisson Gamma Belief Network (PGBN), is proposed to extract multi-layer feature from SAR images data. As a deep Bayesian generative network, PGBN has the ability to extract a multilayer structured representation from the complex SAR images owing to the existence of Poisson likelihood and multilayer gamma hidden variables, at the same time the PGBN can be viewed as a deep non-negative matrix factorization model. Note that the PGBN model is an unsupervised deep generative network and it fails to make full use of the label information in training stage. Therefore, the NBPGBN model is further proposed to obtain a higher recognition performance and training efficiency based on Naïve Bayes rule. The experimental results on MSTAR dataset show that the feature extracted by this new approach has better structured information and perform better classification result compared with some related algorithms.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129412701","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}
Pub Date : 2017-05-01DOI: 10.1109/RSIP.2017.7958813
Yun Zhang, Shiyu Yang, Hongbo Li, Huilin Mu
This paper addresses the problem of track initiation for multi-target in different motion forms and in complicated clutter background. The method proposed combines the traditional logic-based method and convolution neural network. The logic-based method is used mainly to generate a set of track proposals, which is computed by the convolution neural network to extract features in data domain. In this paper, softmax at the end of the convolution neural network is substituted by a one-dimensional two-class classifier for the output layer of the convolution neural network is designed to output a one-dimensional value. There are two key insights in this method: (1) the classification problem has been transformed into target tracking problem on the condition that the set of track proposals is found. (2) the convolution neural network is firstly used in data domain to mine and augment high-level features that make classification more easily. The simulation experiments have shown that this method performs much better than modified Hough transform which is used to initialize tracks traditionally, especially when the targets are maneuver. In the experiments based on real data, this method is proved to be adaptive enough to initialize tracks whose data comes from different radars.
{"title":"A novel multi-target track initiation method based on convolution neural network","authors":"Yun Zhang, Shiyu Yang, Hongbo Li, Huilin Mu","doi":"10.1109/RSIP.2017.7958813","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958813","url":null,"abstract":"This paper addresses the problem of track initiation for multi-target in different motion forms and in complicated clutter background. The method proposed combines the traditional logic-based method and convolution neural network. The logic-based method is used mainly to generate a set of track proposals, which is computed by the convolution neural network to extract features in data domain. In this paper, softmax at the end of the convolution neural network is substituted by a one-dimensional two-class classifier for the output layer of the convolution neural network is designed to output a one-dimensional value. There are two key insights in this method: (1) the classification problem has been transformed into target tracking problem on the condition that the set of track proposals is found. (2) the convolution neural network is firstly used in data domain to mine and augment high-level features that make classification more easily. The simulation experiments have shown that this method performs much better than modified Hough transform which is used to initialize tracks traditionally, especially when the targets are maneuver. In the experiments based on real data, this method is proved to be adaptive enough to initialize tracks whose data comes from different radars.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125702049","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：