Pub Date : 2005-05-16DOI: 10.1109/AMTRSI.2005.1469830
C.A. Collins, D. W. Wilkinson, D. Evans
The use of Landsat data to aid in forest sampling stratification, area estimation, and future resource assessment through growth models is currently being investigated for the state of Mississippi with the goal of better understanding present and future wood resources. In such analyses, and as a part of this investigation, change detection techniques are being exploited to help determine these forest stand ages in approximate five year intervals. This preliminary report looks at post classification comparisons and temporal image differencing as two means to find these dates. The results find the post classification comparisons techniques, in an unrefined use, to work moderately well (overall accuracy = 0.6157, KHAT = 0.5386) and temporal image differencing with NDVI and tasseled cap transformations to disagree with each other in predicted age class sizes with no assessment data to validate accuracy at this time.
{"title":"Multi-temporal analysis of landsat data to determine forest age classes for the mississippi statewide forest inventory~preliminary results","authors":"C.A. Collins, D. W. Wilkinson, D. Evans","doi":"10.1109/AMTRSI.2005.1469830","DOIUrl":"https://doi.org/10.1109/AMTRSI.2005.1469830","url":null,"abstract":"The use of Landsat data to aid in forest sampling stratification, area estimation, and future resource assessment through growth models is currently being investigated for the state of Mississippi with the goal of better understanding present and future wood resources. In such analyses, and as a part of this investigation, change detection techniques are being exploited to help determine these forest stand ages in approximate five year intervals. This preliminary report looks at post classification comparisons and temporal image differencing as two means to find these dates. The results find the post classification comparisons techniques, in an unrefined use, to work moderately well (overall accuracy = 0.6157, KHAT = 0.5386) and temporal image differencing with NDVI and tasseled cap transformations to disagree with each other in predicted age class sizes with no assessment data to validate accuracy at this time.","PeriodicalId":302923,"journal":{"name":"International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, 2005.","volume":"235 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122951795","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 : 2005-05-16DOI: 10.1109/AMTRSI.2005.1469844
G. V. Mostovoy, R. King, K. R. Reddy, V. Kakani
Three datasets of surface and 2-m air temperature (MODIS LST gridded data, North American Regional Reanalysis surface fields, and meteorological observations from surface stations) with different spatial resolution ranging from a field scale to 32-km model grid have been used to evaluate a statistical relationship between Land Surface Temperature (LST) and daily 2-m maximum and minimum air temperature Ta. The datasets cover Mississippi and adjacent states for the period from June 2000 to September 2004. A comparison between correlations produced by MODIS LST versus observed Tmax and Reanalysis LST versus observed Tmax (surface and air temperature) were performed to assess effects of the averaging scale and the diurnal cycle. Seasonal changes in correlation pattern between December, January, and February (DJF) and June, July, and August (JJA) are revealed and examined.
{"title":"Using MODIS LST data for high-resolution estimates of daily air temperature over Mississippi","authors":"G. V. Mostovoy, R. King, K. R. Reddy, V. Kakani","doi":"10.1109/AMTRSI.2005.1469844","DOIUrl":"https://doi.org/10.1109/AMTRSI.2005.1469844","url":null,"abstract":"Three datasets of surface and 2-m air temperature (MODIS LST gridded data, North American Regional Reanalysis surface fields, and meteorological observations from surface stations) with different spatial resolution ranging from a field scale to 32-km model grid have been used to evaluate a statistical relationship between Land Surface Temperature (LST) and daily 2-m maximum and minimum air temperature Ta. The datasets cover Mississippi and adjacent states for the period from June 2000 to September 2004. A comparison between correlations produced by MODIS LST versus observed Tmax and Reanalysis LST versus observed Tmax (surface and air temperature) were performed to assess effects of the averaging scale and the diurnal cycle. Seasonal changes in correlation pattern between December, January, and February (DJF) and June, July, and August (JJA) are revealed and examined.","PeriodicalId":302923,"journal":{"name":"International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, 2005.","volume":"85 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131467938","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 : 2005-05-16DOI: 10.1109/AMTRSI.2005.1469852
M. Pagnutti, K. Holekamp, R. Ryan, R. Vaughan, J. Russell, D. Prados, T. Stanley
Remotely sensed ground reflectance is the basis for many inter-sensor interoperability or change detection techniques. Satellite inter-comparisons and accurate vegetation indices such as the Normalized Difference Vegetation Index, which is used to describe or to imply a wide variety of biophysical parameters and is defined in terms of near-infrared and red- band reflectance, require the generation of accurate reflectance maps. This generation relies upon the removal of solar illumination, satellite geometry, and atmospheric effects and is generally referred to as atmospheric correction. Atmospheric correction of remotely sensed imagery to ground reflectance, however, has been widely applied to only a few systems. In this study, we atmospherically corrected commercially available, high spatial resolution IKONOS and QuickBird imagery using several methods to determine the accuracy of the resulting reflectance maps. We used extensive ground measurement datasets for nine IKONOS and QuickBird scenes acquired over a two-year period to establish reflectance map accuracies. A correction approach using atmospheric products derived from Moderate Resolution Imaging Spectrometer data created excellent reflectance maps and demonstrated a reliable, effective method for reflectance map generation.
{"title":"Atmospheric correction of high spatial resolution commercial satellite imagery products using MODIS atmospheric products","authors":"M. Pagnutti, K. Holekamp, R. Ryan, R. Vaughan, J. Russell, D. Prados, T. Stanley","doi":"10.1109/AMTRSI.2005.1469852","DOIUrl":"https://doi.org/10.1109/AMTRSI.2005.1469852","url":null,"abstract":"Remotely sensed ground reflectance is the basis for many inter-sensor interoperability or change detection techniques. Satellite inter-comparisons and accurate vegetation indices such as the Normalized Difference Vegetation Index, which is used to describe or to imply a wide variety of biophysical parameters and is defined in terms of near-infrared and red- band reflectance, require the generation of accurate reflectance maps. This generation relies upon the removal of solar illumination, satellite geometry, and atmospheric effects and is generally referred to as atmospheric correction. Atmospheric correction of remotely sensed imagery to ground reflectance, however, has been widely applied to only a few systems. In this study, we atmospherically corrected commercially available, high spatial resolution IKONOS and QuickBird imagery using several methods to determine the accuracy of the resulting reflectance maps. We used extensive ground measurement datasets for nine IKONOS and QuickBird scenes acquired over a two-year period to establish reflectance map accuracies. A correction approach using atmospheric products derived from Moderate Resolution Imaging Spectrometer data created excellent reflectance maps and demonstrated a reliable, effective method for reflectance map generation.","PeriodicalId":302923,"journal":{"name":"International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, 2005.","volume":"220 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123444797","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 : 2005-05-16DOI: 10.1109/AMTRSI.2005.1469841
A. Mathur, L. Bruce, J. Madsen
In this paper, the authors investigate the use of hyperspectral-multitemporal features for discriminating between two aquatic weed species, Waterhyacinth and Bulrush. Hyperspectral, multitemporal data is three- dimensional data that can be organized into a "spectro- temporal map" where the x-axis is time, y-axis is wavelength, and z-axis is reflectance. The authors present an algorithm based on a greedy search approach to extract pertinent features to solve the classification problem at hand.
{"title":"Feature extraction via spectro-temporal analysis of hyperspectral data for vegetative target detection","authors":"A. Mathur, L. Bruce, J. Madsen","doi":"10.1109/AMTRSI.2005.1469841","DOIUrl":"https://doi.org/10.1109/AMTRSI.2005.1469841","url":null,"abstract":"In this paper, the authors investigate the use of hyperspectral-multitemporal features for discriminating between two aquatic weed species, Waterhyacinth and Bulrush. Hyperspectral, multitemporal data is three- dimensional data that can be organized into a \"spectro- temporal map\" where the x-axis is time, y-axis is wavelength, and z-axis is reflectance. The authors present an algorithm based on a greedy search approach to extract pertinent features to solve the classification problem at hand.","PeriodicalId":302923,"journal":{"name":"International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, 2005.","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125805797","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 : 2005-05-16DOI: 10.1109/AMTRSI.2005.1469831
L. M. Moskal
Diversity patterns observed for a given landscape are dependent on the spatial and temporal scales of the investigation. In this research the temporal response was investigated within the undisturbed mature forests of the Yellowstone region, the 1988 fire burns and the harvested forest tracks in Targhee National Forest. The satellite data set employed for this analysis was the 1989 to 2001 weekly NOAA AVHRR NDVI composite. Harmonic analysis was used to express the cyclical NDVI curve as a sum of a series of cosine waves and an additive term. Two research questions were addressed: 1) Can interannual and seasonal patterns for the various disturbed and undisturbed forested landscapes be discerned? 2) How do the interannual and seasonal patterns vary for undisturbed, naturally disburdened and human impacted forest landscapes? The changes in amplitude, phase and variance for the most explanatory harmonic term were compared between the two disturbance types and the undisturbed forests. The findings show that interannual and seasonal patterns differ with disturbance type and both differ from undisturbed forests. Mature forests are the most difficult to predict for interannual and seasonal patterns, predictability is easiest for young, post disturbance areas. Finally, the results show that natural forests are temporally
{"title":"Temporal signatures and harmonic analysis of natural and anthropogenic disturbances of forested landscapes: a case study in the Yellowstone region","authors":"L. M. Moskal","doi":"10.1109/AMTRSI.2005.1469831","DOIUrl":"https://doi.org/10.1109/AMTRSI.2005.1469831","url":null,"abstract":"Diversity patterns observed for a given landscape are dependent on the spatial and temporal scales of the investigation. In this research the temporal response was investigated within the undisturbed mature forests of the Yellowstone region, the 1988 fire burns and the harvested forest tracks in Targhee National Forest. The satellite data set employed for this analysis was the 1989 to 2001 weekly NOAA AVHRR NDVI composite. Harmonic analysis was used to express the cyclical NDVI curve as a sum of a series of cosine waves and an additive term. Two research questions were addressed: 1) Can interannual and seasonal patterns for the various disturbed and undisturbed forested landscapes be discerned? 2) How do the interannual and seasonal patterns vary for undisturbed, naturally disburdened and human impacted forest landscapes? The changes in amplitude, phase and variance for the most explanatory harmonic term were compared between the two disturbance types and the undisturbed forests. The findings show that interannual and seasonal patterns differ with disturbance type and both differ from undisturbed forests. Mature forests are the most difficult to predict for interannual and seasonal patterns, predictability is easiest for young, post disturbance areas. Finally, the results show that natural forests are temporally","PeriodicalId":302923,"journal":{"name":"International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, 2005.","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114362831","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 : 2005-05-16DOI: 10.1109/AMTRSI.2005.1469881
G. Nelson, E. Ramsey, A. Rangoonwala
Landsat Thematic Mapper images and collateral data sources were used to classify the land cover of the Mermentau River Basin within the chenier coastal plain and the adjacent uplands of Louisiana, USA. Landcover classes followed that of the National Oceanic and Atmospheric Administration's Coastal Change Analysis Program; however, classification methods needed to be developed to meet these national standards. Our first classification was limited to the Mermentau River Basin (MRB) in southcentral Louisiana, and the years of 1990, 1993, and 1996. To overcome problems due to class spectral inseparable, spatial and spectra continuums, mixed landcovers, and abnormal transitions, we separated the coastal area into regions of commonality and applying masks to specific land mixtures. Over the three years and 14 landcover classes (aggregating the cultivated land and grassland, and water and floating vegetation classes), overall accuracies ranged from 82% to 90%. To enhance landcover change interpretation, three indicators were introduced as Location Stability, Residence stability, and Turnover. Implementing methods substantiated in the multiple date MRB classification, we spatially extended the classification to the entire Louisiana coast and temporally extended the original 1990, 1993, 1996 classifications to 1999 (Figure 1). We also advanced the operational functionality of the classification and increased the credibility of change detection results. Increased operational functionality that resulted in diminished user input was for the most part gained by implementing a classification logic based on forbidden transitions. The logic detected and corrected misclassifications and mostly alleviated the necessity of subregion separation prior to the classification. The new methods provided an improved ability for more timely detection and response to landcover impact.
{"title":"Mapping and improving frequency, accuracy, and interpretation of land cover change: classifying coastal louisiana with 1990, 1993, 1996, and 1999 Landsat thematic mapper image data","authors":"G. Nelson, E. Ramsey, A. Rangoonwala","doi":"10.1109/AMTRSI.2005.1469881","DOIUrl":"https://doi.org/10.1109/AMTRSI.2005.1469881","url":null,"abstract":"Landsat Thematic Mapper images and collateral data sources were used to classify the land cover of the Mermentau River Basin within the chenier coastal plain and the adjacent uplands of Louisiana, USA. Landcover classes followed that of the National Oceanic and Atmospheric Administration's Coastal Change Analysis Program; however, classification methods needed to be developed to meet these national standards. Our first classification was limited to the Mermentau River Basin (MRB) in southcentral Louisiana, and the years of 1990, 1993, and 1996. To overcome problems due to class spectral inseparable, spatial and spectra continuums, mixed landcovers, and abnormal transitions, we separated the coastal area into regions of commonality and applying masks to specific land mixtures. Over the three years and 14 landcover classes (aggregating the cultivated land and grassland, and water and floating vegetation classes), overall accuracies ranged from 82% to 90%. To enhance landcover change interpretation, three indicators were introduced as Location Stability, Residence stability, and Turnover. Implementing methods substantiated in the multiple date MRB classification, we spatially extended the classification to the entire Louisiana coast and temporally extended the original 1990, 1993, 1996 classifications to 1999 (Figure 1). We also advanced the operational functionality of the classification and increased the credibility of change detection results. Increased operational functionality that resulted in diminished user input was for the most part gained by implementing a classification logic based on forbidden transitions. The logic detected and corrected misclassifications and mostly alleviated the necessity of subregion separation prior to the classification. The new methods provided an improved ability for more timely detection and response to landcover impact.","PeriodicalId":302923,"journal":{"name":"International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, 2005.","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133183274","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 : 2005-05-16DOI: 10.1109/AMTRSI.2005.1469860
M. L. Aten, G. Easson
Accelerated eutrophication is often the result of man-induced nutrient enrichment into a water basin as a result of agricultural and industrial practices. This influx of excessive nutrients catalyzes the intense growth of algae and aquatic plants, which has a negative impact on water quality, making it unsuitable for consumption. This research proposes to investigate the potential for satellite imagery to provide continuous monitoring capabilities for advanced notification of lake eutrophication for the management of water resources. We will use in situ field data collected from monitoring stations to calibrate the collected data and to assess the accuracy of this approach. We will utilize historic and current satellite imagery to produce a series of maps for topography, bathymetry, land use, land cover, water clarity, water circulation, water temperature, vegetation indices, and other relevant data layers. This information will permit a comprehensive basin characterization that will facilitate the correlation to field data and the identification of those factors that lead to lake eutrophication.
{"title":"Lake eutrophication change detection for the management of water resources","authors":"M. L. Aten, G. Easson","doi":"10.1109/AMTRSI.2005.1469860","DOIUrl":"https://doi.org/10.1109/AMTRSI.2005.1469860","url":null,"abstract":"Accelerated eutrophication is often the result of man-induced nutrient enrichment into a water basin as a result of agricultural and industrial practices. This influx of excessive nutrients catalyzes the intense growth of algae and aquatic plants, which has a negative impact on water quality, making it unsuitable for consumption. This research proposes to investigate the potential for satellite imagery to provide continuous monitoring capabilities for advanced notification of lake eutrophication for the management of water resources. We will use in situ field data collected from monitoring stations to calibrate the collected data and to assess the accuracy of this approach. We will utilize historic and current satellite imagery to produce a series of maps for topography, bathymetry, land use, land cover, water clarity, water circulation, water temperature, vegetation indices, and other relevant data layers. This information will permit a comprehensive basin characterization that will facilitate the correlation to field data and the identification of those factors that lead to lake eutrophication.","PeriodicalId":302923,"journal":{"name":"International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, 2005.","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124155576","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 : 2005-05-16DOI: 10.1109/AMTRSI.2005.1469829
D. Goodenough, Hao Chen, A. Dyk, T. Han
Multitemporal Envisat ASAR precision images with alternating polarization mode (1) were collected in this study to investigate radar backscatter variability for different forest types in a northern forest environment in Canada. Prior to the analysis, the images were pre-processed, including speckle reduction, SAR texture generation, and image orthorectification. A 2002 Landsat TM image was prepared for the SAR-optical data fusion analysis. To determine the effectiveness of C-band Envisat ASAR data for the use of forest mapping, structure recognition, and change detection, a hierarchical logistic classifier, LOGIT (2), was used to classify the multitemporal and multi-polarization ASAR images. The scattering characteristics of different forest covers, clear-cut areas, and forest regeneration were examined and the classification comparisons were made. This paper reports on these experiments and the methodology for using multi-temporal and multi-polarization SAR in northern forest environments.
{"title":"Evaluation of multi-temporal and multi-polarization ASAR for Boreal Forests in Hinton","authors":"D. Goodenough, Hao Chen, A. Dyk, T. Han","doi":"10.1109/AMTRSI.2005.1469829","DOIUrl":"https://doi.org/10.1109/AMTRSI.2005.1469829","url":null,"abstract":"Multitemporal Envisat ASAR precision images with alternating polarization mode (1) were collected in this study to investigate radar backscatter variability for different forest types in a northern forest environment in Canada. Prior to the analysis, the images were pre-processed, including speckle reduction, SAR texture generation, and image orthorectification. A 2002 Landsat TM image was prepared for the SAR-optical data fusion analysis. To determine the effectiveness of C-band Envisat ASAR data for the use of forest mapping, structure recognition, and change detection, a hierarchical logistic classifier, LOGIT (2), was used to classify the multitemporal and multi-polarization ASAR images. The scattering characteristics of different forest covers, clear-cut areas, and forest regeneration were examined and the classification comparisons were made. This paper reports on these experiments and the methodology for using multi-temporal and multi-polarization SAR in northern forest environments.","PeriodicalId":302923,"journal":{"name":"International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, 2005.","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115902444","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 : 2005-05-16DOI: 10.1109/AMTRSI.2005.1469864
A. Nielsen, M. Canty
Change detection methods for multi- and hyper- variate data aim at identifying differences in data acquired over the same area at different points in time. In this con- tribution an iterative extension to the multivariate alteration detection (MAD) transformation for change detection is sketched and applied. The MAD transformation is based on canonical correlation analysis (CCA), which is an established technique in multivariate statistics. The extension in an iterative scheme seeks to establish an increasingly better background of no-change against which to detect change. This is done by putting higher weights on observations of no-change in the calculation of the statistics for the CCA. The differences found may be due to noise or differences in (atmospheric etc.) conditions at the two acquisition time points. To prevent a change detection method from detecting uninteresting change due to noise or arbitrary spurious differences the application of regularization, also known as penalization, and other types of robustification of the change detection method may be important especially when applied to hyperspectral data. Among other things results show that the new iterated scheme does give a better no-change background against which to detect change than the original, non-iterative MAD method and that the IR-MAD method depicts the change detected in less noisy components. I. INTRODUCTION This contribution focuses on construction of more gen- eral difference images than simple differences in multivariate change detection. This is done via an iterated version (1) of the canonical correlation analysis (CCA) (2) based multivariate alteration detection (MAD) method (3) that could, moreover, be combined with an expectation-maximization (EM) based method for determining thresholds for differentiating between change and no-change in the difference images, and for estimating the variance-covariance structure of the no-change observations (4), (5). The variances can be used to estab- lish a single change/no-change image based on the general multivariate difference image. The resulting imagery from MAD based change detection is invariant to linear and affine transformations of the input including, e.g., affine corrections to normalize data between the two acquisition time points. This is an enormous advantage over other multivariate change detection methods. The resulting single change/no-change image can be used to establish both change regions and to extract observations with which a fully automated orthogonal regression analysis based normalization of the multivariate data between the two points in time can be developed (6). Results (not shown here) from partly simulated multivariate data indicate an improved performance of the iterated scheme over the original MAD method (1). Also, a few comparisons with established methods for calculation of robust statistics for the CCA indicate that the scheme suggested here performs better, see also (7). Regularizati
{"title":"Multi- and hyperspectral remote sensing change detection with generalized difference images by the IR-MAD method","authors":"A. Nielsen, M. Canty","doi":"10.1109/AMTRSI.2005.1469864","DOIUrl":"https://doi.org/10.1109/AMTRSI.2005.1469864","url":null,"abstract":"Change detection methods for multi- and hyper- variate data aim at identifying differences in data acquired over the same area at different points in time. In this con- tribution an iterative extension to the multivariate alteration detection (MAD) transformation for change detection is sketched and applied. The MAD transformation is based on canonical correlation analysis (CCA), which is an established technique in multivariate statistics. The extension in an iterative scheme seeks to establish an increasingly better background of no-change against which to detect change. This is done by putting higher weights on observations of no-change in the calculation of the statistics for the CCA. The differences found may be due to noise or differences in (atmospheric etc.) conditions at the two acquisition time points. To prevent a change detection method from detecting uninteresting change due to noise or arbitrary spurious differences the application of regularization, also known as penalization, and other types of robustification of the change detection method may be important especially when applied to hyperspectral data. Among other things results show that the new iterated scheme does give a better no-change background against which to detect change than the original, non-iterative MAD method and that the IR-MAD method depicts the change detected in less noisy components. I. INTRODUCTION This contribution focuses on construction of more gen- eral difference images than simple differences in multivariate change detection. This is done via an iterated version (1) of the canonical correlation analysis (CCA) (2) based multivariate alteration detection (MAD) method (3) that could, moreover, be combined with an expectation-maximization (EM) based method for determining thresholds for differentiating between change and no-change in the difference images, and for estimating the variance-covariance structure of the no-change observations (4), (5). The variances can be used to estab- lish a single change/no-change image based on the general multivariate difference image. The resulting imagery from MAD based change detection is invariant to linear and affine transformations of the input including, e.g., affine corrections to normalize data between the two acquisition time points. This is an enormous advantage over other multivariate change detection methods. The resulting single change/no-change image can be used to establish both change regions and to extract observations with which a fully automated orthogonal regression analysis based normalization of the multivariate data between the two points in time can be developed (6). Results (not shown here) from partly simulated multivariate data indicate an improved performance of the iterated scheme over the original MAD method (1). Also, a few comparisons with established methods for calculation of robust statistics for the CCA indicate that the scheme suggested here performs better, see also (7). Regularizati","PeriodicalId":302923,"journal":{"name":"International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, 2005.","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122848073","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 : 2005-05-16DOI: 10.1109/AMTRSI.2005.1469862
C. Butson, R. Fraser
As climate change research becomes increasingly concerned with predicting future trends in the net balance of atmosphere and biosphere CO2 , mapping land cover changes using remote sensing imagery may aid in systematic monitoring for these efforts. This is of special interest in northern areas as they may be more susceptible to rapid change, causing migrations of the tree line and altered permafrost depths. In the current study, we examine and quantify various land cover changes from 1975 to 2001 using multi-temporal Landsat imagery over four pilot sites located in northern Canada. To assess land cover change, three change detection methods were tested using a reference land cover map created by spectral clustering of the most current circa 2000 Landsat ETM+ scene. The three methods under comparison were: 1) Cross-correlation Analysis (CCA), 2) Change Vector Analysis (CVA) and 3) Theil-Sen Regression Analysis (TSA). The methods are similar in that they perform cluster-based statistical analysis going back through the historic data available for each site. To compare the change techniques, each method was applied to the overlapping region of two Landsat ETM+ data paths acquired less than 9 days apart. Assuming no change between the two Landsat acquisitions, CCA and CVA produced similar commission errors (%1.2) while the TSA commission error improved to %0.02. The dominant commission errors were found in the grassland land cover class. Extending this change analysis to the four pilot areas, each of the methods produced variable results. The maximum change recorded for Site #1 was 2368km 2 between 2000-1992. Site #2 characterized a maximum change of 2558km 2 . The maximum change calculated for Site #3 located in northern Ontario was 1983km 2 while the site in Quebec changed by 1031km
{"title":"Mapping land cover change and terrestrial dynamics over northern canada using multi-temporal landsat imagery","authors":"C. Butson, R. Fraser","doi":"10.1109/AMTRSI.2005.1469862","DOIUrl":"https://doi.org/10.1109/AMTRSI.2005.1469862","url":null,"abstract":"As climate change research becomes increasingly concerned with predicting future trends in the net balance of atmosphere and biosphere CO2 , mapping land cover changes using remote sensing imagery may aid in systematic monitoring for these efforts. This is of special interest in northern areas as they may be more susceptible to rapid change, causing migrations of the tree line and altered permafrost depths. In the current study, we examine and quantify various land cover changes from 1975 to 2001 using multi-temporal Landsat imagery over four pilot sites located in northern Canada. To assess land cover change, three change detection methods were tested using a reference land cover map created by spectral clustering of the most current circa 2000 Landsat ETM+ scene. The three methods under comparison were: 1) Cross-correlation Analysis (CCA), 2) Change Vector Analysis (CVA) and 3) Theil-Sen Regression Analysis (TSA). The methods are similar in that they perform cluster-based statistical analysis going back through the historic data available for each site. To compare the change techniques, each method was applied to the overlapping region of two Landsat ETM+ data paths acquired less than 9 days apart. Assuming no change between the two Landsat acquisitions, CCA and CVA produced similar commission errors (%1.2) while the TSA commission error improved to %0.02. The dominant commission errors were found in the grassland land cover class. Extending this change analysis to the four pilot areas, each of the methods produced variable results. The maximum change recorded for Site #1 was 2368km 2 between 2000-1992. Site #2 characterized a maximum change of 2558km 2 . The maximum change calculated for Site #3 located in northern Ontario was 1983km 2 while the site in Quebec changed by 1031km","PeriodicalId":302923,"journal":{"name":"International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, 2005.","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131961676","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：