Pub Date : 2003-10-27DOI: 10.1109/WARSD.2003.1295198
M. Graña, J. Gallego, C. Hernández
Our main interest is to perform unsupervised segmentation of the hyperspectral images. Our approach is to interpret abundance images resulting from spectral unmixing as the characterization of regions in the image. We induce the endmembers needed for spectral unmixing from the image data. Therefore the endmember spectra are not easily interpretable as laboratory spectra. Our method for endmember induction looks at the morphological independence or the endmembers as a necessary condition. We use the Autoassociative Morphological Memories (AMM) as detectors of morphological independence conditions. Our algorithm needs only one pass of the image. The experimental results obtained over a set of synthetic images are presented here, contrasted with the ICA and CCA approaches.
{"title":"Further results on AMM for endmember induction","authors":"M. Graña, J. Gallego, C. Hernández","doi":"10.1109/WARSD.2003.1295198","DOIUrl":"https://doi.org/10.1109/WARSD.2003.1295198","url":null,"abstract":"Our main interest is to perform unsupervised segmentation of the hyperspectral images. Our approach is to interpret abundance images resulting from spectral unmixing as the characterization of regions in the image. We induce the endmembers needed for spectral unmixing from the image data. Therefore the endmember spectra are not easily interpretable as laboratory spectra. Our method for endmember induction looks at the morphological independence or the endmembers as a necessary condition. We use the Autoassociative Morphological Memories (AMM) as detectors of morphological independence conditions. Our algorithm needs only one pass of the image. The experimental results obtained over a set of synthetic images are presented here, contrasted with the ICA and CCA approaches.","PeriodicalId":395735,"journal":{"name":"IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, 2003","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117092492","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 : 2003-10-27DOI: 10.1109/WARSD.2003.1295210
C. Boehm, R. Schenkel
Parallel to the conventional (statistical, spectral) description of mixed urban classes for image segmentation, the description on the basis of cues and related spatial properties is used within the classification process. Recently we concentrate very much on strong model-based classification, which may lead to a classification not covering the whole area due to the implementation of insufficient models (class descriptions). Major interest is related to urban features like urban fabric, continuous urban fabric (dense, medium dense), discontinuous urban fabric (dense residential, sparse residential, residential blocks) as well as industrial areas.
{"title":"Analysis of high resolution polarimetric SAR in urban areas","authors":"C. Boehm, R. Schenkel","doi":"10.1109/WARSD.2003.1295210","DOIUrl":"https://doi.org/10.1109/WARSD.2003.1295210","url":null,"abstract":"Parallel to the conventional (statistical, spectral) description of mixed urban classes for image segmentation, the description on the basis of cues and related spatial properties is used within the classification process. Recently we concentrate very much on strong model-based classification, which may lead to a classification not covering the whole area due to the implementation of insufficient models (class descriptions). Major interest is related to urban features like urban fabric, continuous urban fabric (dense, medium dense), discontinuous urban fabric (dense residential, sparse residential, residential blocks) as well as industrial areas.","PeriodicalId":395735,"journal":{"name":"IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, 2003","volume":"111 3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129076103","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 : 2003-10-27DOI: 10.1109/WARSD.2003.1295209
F. Dell’acqua, P. Gamba
In this work we improve a methodology for discriminating urban environments by means of textural features in SAR images. In particular, we introduce multi-scale co-occurrence features and show how the feature set may be chosen as a function of the training set and the mapping classes. Moreover, we provide and compare results obtained by different satellite SAR sensors on the same urban test site, as well as a combination of these sets. Finally, a short analysis of the polarization effects and their importance in this framework of analysis is considered. The results are extremely encouraging, and show the potential of this technique, even if more research is needed to exploit the capabilities of the new generation of low-Earth orbit SAR satellites.
{"title":"Discriminating urban environments using multi-scale texture and multiple SAR images","authors":"F. Dell’acqua, P. Gamba","doi":"10.1109/WARSD.2003.1295209","DOIUrl":"https://doi.org/10.1109/WARSD.2003.1295209","url":null,"abstract":"In this work we improve a methodology for discriminating urban environments by means of textural features in SAR images. In particular, we introduce multi-scale co-occurrence features and show how the feature set may be chosen as a function of the training set and the mapping classes. Moreover, we provide and compare results obtained by different satellite SAR sensors on the same urban test site, as well as a combination of these sets. Finally, a short analysis of the polarization effects and their importance in this framework of analysis is considered. The results are extremely encouraging, and show the potential of this technique, even if more research is needed to exploit the capabilities of the new generation of low-Earth orbit SAR satellites.","PeriodicalId":395735,"journal":{"name":"IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, 2003","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130766169","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 : 2003-10-27DOI: 10.1109/WARSD.2003.1295208
Antonio Plaza, P. Martínez, J. Plaza, R. Pérez
The integration of spatial and spectral responses in hyperspectral image data analysis has been identified as a desirable objective by the remote sensing community. However, most available attempts are based on the consideration of spectral information separately from spatial information, and thus the two types of information are not treated simultaneously. In this paper, we describe our background in applying joint spatial/spectral techniques for full (pure)- and mixed-pixel classification of hyperspectral image data. Most of the techniques described in this work are based on classic mathematical morphology theory, which provides a remarkable framework to achieve the desired integration. The performance of the proposed methodologies is demonstrated by comparing them to other well-known pure- and mixed-pixel classifiers, using both simulated and real hyperspectral data collected by the NASA/JPL-AVIRIS and DLR-DAIS 7915 imaging spectrometers.
{"title":"Spatial/Spectral analysis of hyperspectral image data","authors":"Antonio Plaza, P. Martínez, J. Plaza, R. Pérez","doi":"10.1109/WARSD.2003.1295208","DOIUrl":"https://doi.org/10.1109/WARSD.2003.1295208","url":null,"abstract":"The integration of spatial and spectral responses in hyperspectral image data analysis has been identified as a desirable objective by the remote sensing community. However, most available attempts are based on the consideration of spectral information separately from spatial information, and thus the two types of information are not treated simultaneously. In this paper, we describe our background in applying joint spatial/spectral techniques for full (pure)- and mixed-pixel classification of hyperspectral image data. Most of the techniques described in this work are based on classic mathematical morphology theory, which provides a remarkable framework to achieve the desired integration. The performance of the proposed methodologies is demonstrated by comparing them to other well-known pure- and mixed-pixel classifiers, using both simulated and real hyperspectral data collected by the NASA/JPL-AVIRIS and DLR-DAIS 7915 imaging spectrometers.","PeriodicalId":395735,"journal":{"name":"IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, 2003","volume":"289 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117293498","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 : 2003-10-27DOI: 10.1109/WARSD.2003.1295221
A. Cheriyadat, L. Bruce, A. Mathur
In recent years, more intuitive understanding about the characteristics of higher dimensional space has influenced the development of subsequent data analysis and classification algorithms in the field of hyperspectral remote sensing. Earlier data analysis and classification algorithms rely on processing high dimensional space as a whole to extract a lower dimensional feature space. The major impediment on these techniques is the limited training data size, which does not confer with the large dimensionality of hyperspectral data. Previous work has shown that statistically reliable parameter estimation can be performed on lower dimensional subspaces that are formed by decomposing the entire dimension into a set of subspaces (bases), based on certain discrimination criterion. In this paper the authors present a classification technique that combines the feature level fusion capabilities of lower dimensional subspaces; with decision level fusion to improve the classification potential of hyperspectral data. In order to reduce the impact of conflicting decisions by individual bases, a voting scheme called Qualified Majority Voting (QMV) is used in combining the decisions. Each base is qualified to influence the final decision, based on its ability to predict the classes with respect to other bases. This information can be derived from training data, analyst inputs or feed back from prior applications. Unlike the traditional classification approaches, this technique not only utilizes the projected lower dimensional feature space, but also makes use of the reliability of the subspaces in classifying certain classes.
{"title":"Decision level fusion with best-bases for hyperspectral classification","authors":"A. Cheriyadat, L. Bruce, A. Mathur","doi":"10.1109/WARSD.2003.1295221","DOIUrl":"https://doi.org/10.1109/WARSD.2003.1295221","url":null,"abstract":"In recent years, more intuitive understanding about the characteristics of higher dimensional space has influenced the development of subsequent data analysis and classification algorithms in the field of hyperspectral remote sensing. Earlier data analysis and classification algorithms rely on processing high dimensional space as a whole to extract a lower dimensional feature space. The major impediment on these techniques is the limited training data size, which does not confer with the large dimensionality of hyperspectral data. Previous work has shown that statistically reliable parameter estimation can be performed on lower dimensional subspaces that are formed by decomposing the entire dimension into a set of subspaces (bases), based on certain discrimination criterion. In this paper the authors present a classification technique that combines the feature level fusion capabilities of lower dimensional subspaces; with decision level fusion to improve the classification potential of hyperspectral data. In order to reduce the impact of conflicting decisions by individual bases, a voting scheme called Qualified Majority Voting (QMV) is used in combining the decisions. Each base is qualified to influence the final decision, based on its ability to predict the classes with respect to other bases. This information can be derived from training data, analyst inputs or feed back from prior applications. Unlike the traditional classification approaches, this technique not only utilizes the projected lower dimensional feature space, but also makes use of the reliability of the subspaces in classifying certain classes.","PeriodicalId":395735,"journal":{"name":"IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, 2003","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124755190","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 : 2003-10-27DOI: 10.1109/WARSD.2003.1295169
G. Wilkinson
The long term trend in the accuracy of remotely sensed image classification has been investigated using reported results in the journal Photogrammetric Engineering and Remote Sensing in the period since 1989. The results indicate no significant improvement in the performance of classification methodologies over this period. Average classification performance across all results was found to be 72.7% with the average Kappa value being 0.64. Results also indicate no significant correlation between classification performance and number of classes. A good correlation is found between overall percentage accuracy figures and the Kappa coefficient indicating the suitability of either to categorize overall mapping performance. Only a small percentage of papers (8%) were found to provide all background information necessary to make a sophisticated inter-comparison of methods.
{"title":"Are remotely sensed image classification techniques improving ? Results of a long term trend analysis","authors":"G. Wilkinson","doi":"10.1109/WARSD.2003.1295169","DOIUrl":"https://doi.org/10.1109/WARSD.2003.1295169","url":null,"abstract":"The long term trend in the accuracy of remotely sensed image classification has been investigated using reported results in the journal Photogrammetric Engineering and Remote Sensing in the period since 1989. The results indicate no significant improvement in the performance of classification methodologies over this period. Average classification performance across all results was found to be 72.7% with the average Kappa value being 0.64. Results also indicate no significant correlation between classification performance and number of classes. A good correlation is found between overall percentage accuracy figures and the Kappa coefficient indicating the suitability of either to categorize overall mapping performance. Only a small percentage of papers (8%) were found to provide all background information necessary to make a sophisticated inter-comparison of methods.","PeriodicalId":395735,"journal":{"name":"IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, 2003","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130756383","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 : 2003-10-27DOI: 10.1109/WARSD.2003.1295173
James C. Tilton
Describes an approach for producing high quality hierarchically related image segmentations and some first steps towards exploiting the information content of the segmentation hierarchy. Hierarchically related image segmentations are a set of image segmentations at different levels of detail in which the less detailed segmentations can be produced from specific merges of regions contained in the more detailed segmentations. After a general overview of other approaches to image segmentation, the Hierarchical Segmentation (HSEG) algorithm is presented, along with its recursive formulation (RHSEG). Finally, an approach is outlined for exploiting the information content from the segmentation hierarchy based on changes in region features from one hierarchical level to the next. Comparative results are presented with Landsat Thematic Mapper (TM) data.
{"title":"Analysis of hierarchically related image segmentations","authors":"James C. Tilton","doi":"10.1109/WARSD.2003.1295173","DOIUrl":"https://doi.org/10.1109/WARSD.2003.1295173","url":null,"abstract":"Describes an approach for producing high quality hierarchically related image segmentations and some first steps towards exploiting the information content of the segmentation hierarchy. Hierarchically related image segmentations are a set of image segmentations at different levels of detail in which the less detailed segmentations can be produced from specific merges of regions contained in the more detailed segmentations. After a general overview of other approaches to image segmentation, the Hierarchical Segmentation (HSEG) algorithm is presented, along with its recursive formulation (RHSEG). Finally, an approach is outlined for exploiting the information content from the segmentation hierarchy based on changes in region features from one hierarchical level to the next. Comparative results are presented with Landsat Thematic Mapper (TM) data.","PeriodicalId":395735,"journal":{"name":"IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, 2003","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128836317","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 : 2003-10-27DOI: 10.1109/WARSD.2003.1295177
Xue Liu, M. Kafatos, R. Gomez, S. Goetz
Accurate and reliable information about land cover and land use is essential to carbon cycle and climate change modeling. While historical regional-to-global scale land cover and land use data products had been produced by AVHRR and MSS/TM, this task has been advanced by sensors such as MODIS and ETM since the latter 1990s. While the accuracies and reliabilities of these data products have been improved, there have been reports from the modeling community that additional work is needed to reduce errors so that the uncertainties associated with the global carbon cycle and climate change modeling can be addressed. Remotely sensed data collected in different wavelength regions, at different viewing geometries, usually provide complementary information. Their combination has the potential to enhance remote sensing capabilities in discriminating important land cover components. In this paper, we studied multi-angle data fusion, and optical-SAR data fusion for land cover classification at regional spatial scale in the temperate forests of the eastern United States. Data from EOS-MISR, Landsat-ETM+ and RadarSat-SAR were used. The results showed significantly improved land cover classification accuracy when using the data fusion approach. These results may benefit future land cover products for global change research.
{"title":"Combining MISR, ETM+ and SAR data to improve land cover and land use classification for carbon cycle research","authors":"Xue Liu, M. Kafatos, R. Gomez, S. Goetz","doi":"10.1109/WARSD.2003.1295177","DOIUrl":"https://doi.org/10.1109/WARSD.2003.1295177","url":null,"abstract":"Accurate and reliable information about land cover and land use is essential to carbon cycle and climate change modeling. While historical regional-to-global scale land cover and land use data products had been produced by AVHRR and MSS/TM, this task has been advanced by sensors such as MODIS and ETM since the latter 1990s. While the accuracies and reliabilities of these data products have been improved, there have been reports from the modeling community that additional work is needed to reduce errors so that the uncertainties associated with the global carbon cycle and climate change modeling can be addressed. Remotely sensed data collected in different wavelength regions, at different viewing geometries, usually provide complementary information. Their combination has the potential to enhance remote sensing capabilities in discriminating important land cover components. In this paper, we studied multi-angle data fusion, and optical-SAR data fusion for land cover classification at regional spatial scale in the temperate forests of the eastern United States. Data from EOS-MISR, Landsat-ETM+ and RadarSat-SAR were used. The results showed significantly improved land cover classification accuracy when using the data fusion approach. These results may benefit future land cover products for global change research.","PeriodicalId":395735,"journal":{"name":"IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, 2003","volume":"470 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126213433","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 : 2003-10-27DOI: 10.1109/WARSD.2003.1295225
V. Salomonson, R. Wolfe
The Moderate Resolution Imaging Spectroradiometer (MODIS) is on the NASA Earth Observing System (EOS) Terra and Aqua satellites. The MODIS geolocation approach operationally characterizes MODIS geolocation errors and enables individual MODIS observations to be geolocated to the sub-pixel accuracies required for terrestrial global change applications. An overview of the approach, results from both missions and future work are described.
{"title":"MODIS geolocation approach, results and the future","authors":"V. Salomonson, R. Wolfe","doi":"10.1109/WARSD.2003.1295225","DOIUrl":"https://doi.org/10.1109/WARSD.2003.1295225","url":null,"abstract":"The Moderate Resolution Imaging Spectroradiometer (MODIS) is on the NASA Earth Observing System (EOS) Terra and Aqua satellites. The MODIS geolocation approach operationally characterizes MODIS geolocation errors and enables individual MODIS observations to be geolocated to the sub-pixel accuracies required for terrestrial global change applications. An overview of the approach, results from both missions and future work are described.","PeriodicalId":395735,"journal":{"name":"IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, 2003","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126519921","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 : 2003-10-27DOI: 10.1109/WARSD.2003.1295185
C. K. BrewerA, James A. BarberA, Gregor WillhauckB, U. Benzb
Forest managers need consistent and continuous data on existing vegetation and landcover to address most land management issues and concerns. The current operational approach used by the USDA Forest Service, Northern Region to produce such data using a multi-source and multi-classifier system is described. The methodological components of this system include: (a) ecogeographic stratification, (b) production of image objects through image segmentation, (c) incorporation of multi-temporal image data and change detection, (d) extensive use of ecological modeling and other ancillary data, (e) generation of reference data integrating field sampled inventory data through a structured aerial photo interpretation process, and (f) utilization of multiple classifiers for different levels of the classification hierarchy.
{"title":"Multi-source and multi-classifier system for regional landcover mapping","authors":"C. K. BrewerA, James A. BarberA, Gregor WillhauckB, U. Benzb","doi":"10.1109/WARSD.2003.1295185","DOIUrl":"https://doi.org/10.1109/WARSD.2003.1295185","url":null,"abstract":"Forest managers need consistent and continuous data on existing vegetation and landcover to address most land management issues and concerns. The current operational approach used by the USDA Forest Service, Northern Region to produce such data using a multi-source and multi-classifier system is described. The methodological components of this system include: (a) ecogeographic stratification, (b) production of image objects through image segmentation, (c) incorporation of multi-temporal image data and change detection, (d) extensive use of ecological modeling and other ancillary data, (e) generation of reference data integrating field sampled inventory data through a structured aerial photo interpretation process, and (f) utilization of multiple classifiers for different levels of the classification hierarchy.","PeriodicalId":395735,"journal":{"name":"IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, 2003","volume":"134 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123783833","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：