Pub Date : 2017-05-18DOI: 10.1109/RSIP.2017.7958794
Yancong Wei, Q. Yuan
We proposed a deep convolutional network for multi-spectral image pan-sharpening to overcome the drawbacks of traditional methods and improve the fusion accuracy. To break the performance limitation of deep networks, residual learning with specific adaption to image fusion tasks is applied to optimize the architecture of proposed network. Results of adequate experiments support that our model can yield high resolution multi-spectral images with state-of-the-art qualities, as the information in both spatial and spectral domains has been accurately preserved.
{"title":"Deep residual learning for remote sensed imagery pansharpening","authors":"Yancong Wei, Q. Yuan","doi":"10.1109/RSIP.2017.7958794","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958794","url":null,"abstract":"We proposed a deep convolutional network for multi-spectral image pan-sharpening to overcome the drawbacks of traditional methods and improve the fusion accuracy. To break the performance limitation of deep networks, residual learning with specific adaption to image fusion tasks is applied to optimize the architecture of proposed network. Results of adequate experiments support that our model can yield high resolution multi-spectral images with state-of-the-art qualities, as the information in both spatial and spectral domains has been accurately preserved.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"99 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124104297","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 : 2017-05-18DOI: 10.1109/RSIP.2017.7958806
Yang Liu, Miaohui Zhang, Pengtao Xu, Zhengyang Guo
Reliable automatic ship detection in Synthetic Aperture Radar (SAR) imagery plays an important role in the surveillance of maritime activity. Apart from the well-known Spectral Residual (SR) and CFAR detector, there has emerged a novel method for SAR ship detection, based on the deep learning features. Within this paper, we present a framework of Sea-Land Segmentation-based Convolutional Neural Network (SLS-CNN) for ship detection that attempts to combine the SLS-CNN detector, saliency computation and corner features. For this, sea-land segmentation based on the heat map of SR saliency and probability distribution of the corner is applied, which is followed by SLS-CNN detector, and a final merged minimum bounding rectangles. The framework has been tested and assessed on ALOS PALSAR and TerraSAR-X imagery. Experimental results on representative SAR images of different kinds of ships demonstrate the efficiency and robustness of our proposed SLS-CNN detector.
{"title":"SAR ship detection using sea-land segmentation-based convolutional neural network","authors":"Yang Liu, Miaohui Zhang, Pengtao Xu, Zhengyang Guo","doi":"10.1109/RSIP.2017.7958806","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958806","url":null,"abstract":"Reliable automatic ship detection in Synthetic Aperture Radar (SAR) imagery plays an important role in the surveillance of maritime activity. Apart from the well-known Spectral Residual (SR) and CFAR detector, there has emerged a novel method for SAR ship detection, based on the deep learning features. Within this paper, we present a framework of Sea-Land Segmentation-based Convolutional Neural Network (SLS-CNN) for ship detection that attempts to combine the SLS-CNN detector, saliency computation and corner features. For this, sea-land segmentation based on the heat map of SR saliency and probability distribution of the corner is applied, which is followed by SLS-CNN detector, and a final merged minimum bounding rectangles. The framework has been tested and assessed on ALOS PALSAR and TerraSAR-X imagery. Experimental results on representative SAR images of different kinds of ships demonstrate the efficiency and robustness of our proposed SLS-CNN detector.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124443969","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 : 2017-05-18DOI: 10.1109/RSIP.2017.7958792
Xiaoyi Shen, Jie Zhang, J. Meng, C. Ke
Sea ice type is the most sensitive variables in Arctic ice monitoring and its detailed information is essential for ice situation evaluation, climate prediction and vessels navigating. In this study, we analyzed the different sea ice types with the Cryosat-2 (CS-2) SAR mode waveform data. The waveform of CS-2 data was describe by a set of parameters: pulse peakiness (PP), leading-edge width (LeW), trailing-edge width (TeW), stack standard deviation (SSD) and Maximum value of the echo waveform (Max)] and backscatter coefficient (Sigma0). Random forest (RF) classifier was chosen to classify ice type and the classification results were compared with Arctic and Antarctic Research Institute (AARI) operational ice charts. The results show that 85% of the Arctic surface type can be correctly classified from November 2015 to May 2016, 83% of the FYI can be correctly identified which is the domain ice type in Arctic. In comparison with Bayesian and K nearest-neighbor classifiers, the classification accuracy of RF increased by 5% and 3% respectively.
海冰类型是北极海冰监测中最敏感的变量，其详细信息对冰情评估、气候预报和船舶航行至关重要。在本研究中，我们利用Cryosat-2 (CS-2) SAR模式波形数据分析了不同海冰类型。CS-2数据的波形由脉冲峰值(PP)、前缘宽度(LeW)、尾缘宽度(TeW)、叠加标准差(SSD)、回波波形最大值(Max)和后向散射系数(Sigma0)等参数来描述。采用随机森林(Random forest, RF)分类器对冰型进行分类，并将分类结果与北极南极研究所(Arctic and Antarctic Research Institute, AARI)的业务冰图进行比较。结果表明，2015年11月至2016年5月，85%的北极地表类型可以正确分类，83%的FYI可以正确识别，这是北极的域冰类型。与贝叶斯和K近邻分类器相比，RF的分类准确率分别提高了5%和3%。
{"title":"Sea ice type classification based on random forest machine learning with Cryosat-2 altimeter data","authors":"Xiaoyi Shen, Jie Zhang, J. Meng, C. Ke","doi":"10.1109/RSIP.2017.7958792","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958792","url":null,"abstract":"Sea ice type is the most sensitive variables in Arctic ice monitoring and its detailed information is essential for ice situation evaluation, climate prediction and vessels navigating. In this study, we analyzed the different sea ice types with the Cryosat-2 (CS-2) SAR mode waveform data. The waveform of CS-2 data was describe by a set of parameters: pulse peakiness (PP), leading-edge width (LeW), trailing-edge width (TeW), stack standard deviation (SSD) and Maximum value of the echo waveform (Max)] and backscatter coefficient (Sigma0). Random forest (RF) classifier was chosen to classify ice type and the classification results were compared with Arctic and Antarctic Research Institute (AARI) operational ice charts. The results show that 85% of the Arctic surface type can be correctly classified from November 2015 to May 2016, 83% of the FYI can be correctly identified which is the domain ice type in Arctic. In comparison with Bayesian and K nearest-neighbor classifiers, the classification accuracy of RF increased by 5% and 3% respectively.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"73 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128373591","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 : 2017-05-18DOI: 10.1109/RSIP.2017.7958809
Zhong Hu, Kun Gao, Zeyang Dou
The subpixel mapping technique can obtain a fine-resolution map of target classes in the hyperspectral remote sensing image based on the spatial dependence. In recent years, the subpixel mapping methods based on Maximum A Posterior framework and Total Variation prior (MAP-TV) has received extensive attention because of its unified framework. However, due to the inherent nonlinearity of the TV prior, the traditional gradient descent algorithm to minimize MAP-TV model is inefficient. In this paper, we propose a fast algorithm to solve the MAP-TV model, which combined the fast iterative shrinkage thresholding algorithm and split Bregman algorithm together. The proposed algorithm split the original problem into several sub-problems, each sub-problem has the closed-form solution and is fast to compute. The numerical experiments reveal that the proposed algorithm is faster than the traditional methods and is suitable for the hyperspectral subpixel mapping applications.
{"title":"A fast hyperspectral subpixel mapping algorithm based on MAP-TV framework","authors":"Zhong Hu, Kun Gao, Zeyang Dou","doi":"10.1109/RSIP.2017.7958809","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958809","url":null,"abstract":"The subpixel mapping technique can obtain a fine-resolution map of target classes in the hyperspectral remote sensing image based on the spatial dependence. In recent years, the subpixel mapping methods based on Maximum A Posterior framework and Total Variation prior (MAP-TV) has received extensive attention because of its unified framework. However, due to the inherent nonlinearity of the TV prior, the traditional gradient descent algorithm to minimize MAP-TV model is inefficient. In this paper, we propose a fast algorithm to solve the MAP-TV model, which combined the fast iterative shrinkage thresholding algorithm and split Bregman algorithm together. The proposed algorithm split the original problem into several sub-problems, each sub-problem has the closed-form solution and is fast to compute. The numerical experiments reveal that the proposed algorithm is faster than the traditional methods and is suitable for the hyperspectral subpixel mapping applications.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"85 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134032029","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 : 2017-05-18DOI: 10.1109/RSIP.2017.7958796
W. Xia, Nan. Zhong, Danyang Geng, L. Luo
Extracting road information from remote sensing images plays an import role for many practical areas. In this paper, an approach for road extraction is proposed, in order to obtain standard road region with high accuracy. By utilizing the road design and construction specifications made by the transportation industry, the road objects are assigned into different classes. Then the corresponding task is considered as an image segmentation approach, and deep convolutional network is applied to perform pixel-level estimation to predict the ownership probability of different classes. Besides, a modification processing approach is presented to exploit the segmentation result and obtain formal road network by connecting the missing or unsmooth road subsections. Experiments on remote sensing images are performed, and show that the method is efficient for acquiring multi-type roads from complex situations.
{"title":"A weakly supervised road extraction approach via deep convolutional nets based image segmentation","authors":"W. Xia, Nan. Zhong, Danyang Geng, L. Luo","doi":"10.1109/RSIP.2017.7958796","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958796","url":null,"abstract":"Extracting road information from remote sensing images plays an import role for many practical areas. In this paper, an approach for road extraction is proposed, in order to obtain standard road region with high accuracy. By utilizing the road design and construction specifications made by the transportation industry, the road objects are assigned into different classes. Then the corresponding task is considered as an image segmentation approach, and deep convolutional network is applied to perform pixel-level estimation to predict the ownership probability of different classes. Besides, a modification processing approach is presented to exploit the segmentation result and obtain formal road network by connecting the missing or unsmooth road subsections. Experiments on remote sensing images are performed, and show that the method is efficient for acquiring multi-type roads from complex situations.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"200 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116151519","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 : 2017-05-18DOI: 10.1109/RSIP.2017.7958807
Yizhou Rao, Lin He, Jiawei Zhu
Pan-sharpening has become an important tool in remote sensing, which normally aims at fusing a multi-spectral image with high spectral resolution and a panchromatic image with high spatial resolution. However, some problems, such as spectral distortion, are facing pan-sharpening methods. Inspired by the applications of convolutional neural network (CNN) in many areas, we adopt an effective CNN model to fulfill pan-sharpening. In our method, only the sparse residuals between the interpolated MS and the pan-sharpened image are learned, which achieves fast convergence and high pan-sharpening quality. The experimental results on real-world data validate the effectiveness of the method.
{"title":"A residual convolutional neural network for pan-shaprening","authors":"Yizhou Rao, Lin He, Jiawei Zhu","doi":"10.1109/RSIP.2017.7958807","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958807","url":null,"abstract":"Pan-sharpening has become an important tool in remote sensing, which normally aims at fusing a multi-spectral image with high spectral resolution and a panchromatic image with high spatial resolution. However, some problems, such as spectral distortion, are facing pan-sharpening methods. Inspired by the applications of convolutional neural network (CNN) in many areas, we adopt an effective CNN model to fulfill pan-sharpening. In our method, only the sparse residuals between the interpolated MS and the pan-sharpened image are learned, which achieves fast convergence and high pan-sharpening quality. The experimental results on real-world data validate the effectiveness of the method.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"108 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125008294","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 : 2017-05-18DOI: 10.1109/RSIP.2017.7958811
Jinxin Li, Chao Wang, Shigang Wang, Hong Zhang, Bo Zhang
The new advanced very high resolution (VHR) synthetic aperture radar (SAR) sensor is a kind of high-tech imaging radar developed rapidly in recent years, and it can get even less than 1 m high resolution SAR image. The feature of the VHR SAR image is different from the low or medium resolution SAR image and it contains more abundant information, so the traditional SAR image classification methods can't be directly applied in VHR SAR image classification. In order to achieve high precision classification performance of the VHR SAR image, convolutional neural network (CNN), a kind of representative deep learning method, is applied in this paper. Compared with the traditional supervised classification methods, such as minimum distance and maximum likelihood, the CNN method obtained better classification result with 97.0% average accuracy. The experiments demonstrate that the CNN is an effective and favorable classification method for VHR SAR image classification.
{"title":"Classification of very high resolution SAR image based on convolutional neural network","authors":"Jinxin Li, Chao Wang, Shigang Wang, Hong Zhang, Bo Zhang","doi":"10.1109/RSIP.2017.7958811","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958811","url":null,"abstract":"The new advanced very high resolution (VHR) synthetic aperture radar (SAR) sensor is a kind of high-tech imaging radar developed rapidly in recent years, and it can get even less than 1 m high resolution SAR image. The feature of the VHR SAR image is different from the low or medium resolution SAR image and it contains more abundant information, so the traditional SAR image classification methods can't be directly applied in VHR SAR image classification. In order to achieve high precision classification performance of the VHR SAR image, convolutional neural network (CNN), a kind of representative deep learning method, is applied in this paper. Compared with the traditional supervised classification methods, such as minimum distance and maximum likelihood, the CNN method obtained better classification result with 97.0% average accuracy. The experiments demonstrate that the CNN is an effective and favorable classification method for VHR SAR image classification.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"23 4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126187941","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 : 2017-05-18DOI: 10.1109/RSIP.2017.7958803
Wang Changjie, Nian Hua
Feature extraction is an important method to obtain remote sensing image information. It has significant influence on recognition, analysis, matching, fusion, segmentation of remote sensing image. Image registration is usually classified into two categories: the intensity-based method and the feature-based method. The corner-point is one of the basic features of the images, which has many information and can easily be detected. In the area of remote sensing image application, two or more images are usually mosaiced as one image. According to remote sensing image matching, a method of image matching based on Harris corner-point combined with SURF algorithm is proposed in this paper. First of all, feature points are detected using Harris algorithm, that has the ability of noise immunity and stability. Then, calculating by SURF algorithm, the main directions of the feature points are determined and the feature descriptors are generated. Ratio method is used to get initial matching, and RANSAC algorithm is used to eliminate errors and achieve accurate matching. The result demonstrates that the Harris corner-point image registration described is stable and efficient. The method can be well applied in the remote sensing image processing and geometric positioning accuracy evaluation.
{"title":"Algorithm of remote sensing image matching based on corner-point","authors":"Wang Changjie, Nian Hua","doi":"10.1109/RSIP.2017.7958803","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958803","url":null,"abstract":"Feature extraction is an important method to obtain remote sensing image information. It has significant influence on recognition, analysis, matching, fusion, segmentation of remote sensing image. Image registration is usually classified into two categories: the intensity-based method and the feature-based method. The corner-point is one of the basic features of the images, which has many information and can easily be detected. In the area of remote sensing image application, two or more images are usually mosaiced as one image. According to remote sensing image matching, a method of image matching based on Harris corner-point combined with SURF algorithm is proposed in this paper. First of all, feature points are detected using Harris algorithm, that has the ability of noise immunity and stability. Then, calculating by SURF algorithm, the main directions of the feature points are determined and the feature descriptors are generated. Ratio method is used to get initial matching, and RANSAC algorithm is used to eliminate errors and achieve accurate matching. The result demonstrates that the Harris corner-point image registration described is stable and efficient. The method can be well applied in the remote sensing image processing and geometric positioning accuracy evaluation.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"74 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115234551","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.2017.7958793
Xintong Tan, Jubo Zhu
This paper aims at the feature enhancement for multi-polarimetric synthetic aperture radar (SAR) images. A novel approach based on PDE and regularization which is an extension of the original PDE and regularization methods is proposed. It contains the PDE term for speckle suppression and the sparsity constraint term for strong scatter enhancement. The PDE term is established by combining the ROA detected operator and the amplitude of the multi-polarimetric SAR images. The sparsity constraint term contains the structural information and the sparsity of the images. Experiments on the measured multi-polarimetric SAR images show that the proposed approach can efficiently suppress speckle noise and enhance features especially structural and edge features in SAR images.
{"title":"Feature enhancement for multi-polarimetric SAR images: A novel approach based on PDE and regularization","authors":"Xintong Tan, Jubo Zhu","doi":"10.1109/RSIP.2017.7958793","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958793","url":null,"abstract":"This paper aims at the feature enhancement for multi-polarimetric synthetic aperture radar (SAR) images. A novel approach based on PDE and regularization which is an extension of the original PDE and regularization methods is proposed. It contains the PDE term for speckle suppression and the sparsity constraint term for strong scatter enhancement. The PDE term is established by combining the ROA detected operator and the amplitude of the multi-polarimetric SAR images. The sparsity constraint term contains the structural information and the sparsity of the images. Experiments on the measured multi-polarimetric SAR images show that the proposed approach can efficiently suppress speckle noise and enhance features especially structural and edge features in SAR images.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"237 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":"121630134","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.2017.7958817
Sicong Liu, Yangdong Li, X. Tong
This paper presents a novel unsupervised superpixel-based change detection approach to detect multiple changes in Very-High-Resolution remote sensing images. The proposed approach investigates the spectral-spatial variations at superpixel level which aims to enhance the traditional pixel level change detection performance. In particular, superpixel representation of the spectral change vectors is built by exploiting the homogeneity of local objects associating with the change and no-change classes. A decision-level ensemble strategy is designed to generate a reliable binary change detection result. Then the multi-class changes are identified by automatic clustering. Sensitivity of the relevant parameters are analyzed and discussed. Experimental results obtained on a pair of real VHR images confirm the effectiveness of the proposed approach.
{"title":"Superpixel-based multiple change detection in very-high-resolution remote sensing images","authors":"Sicong Liu, Yangdong Li, X. Tong","doi":"10.1109/RSIP.2017.7958817","DOIUrl":"https://doi.org/10.1109/RSIP.2017.7958817","url":null,"abstract":"This paper presents a novel unsupervised superpixel-based change detection approach to detect multiple changes in Very-High-Resolution remote sensing images. The proposed approach investigates the spectral-spatial variations at superpixel level which aims to enhance the traditional pixel level change detection performance. In particular, superpixel representation of the spectral change vectors is built by exploiting the homogeneity of local objects associating with the change and no-change classes. A decision-level ensemble strategy is designed to generate a reliable binary change detection result. Then the multi-class changes are identified by automatic clustering. Sensitivity of the relevant parameters are analyzed and discussed. Experimental results obtained on a pair of real VHR images confirm the effectiveness of the proposed approach.","PeriodicalId":262222,"journal":{"name":"2017 International Workshop on Remote Sensing with Intelligent Processing (RSIP)","volume":"69 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":"134406630","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：