A comparison of gap-filling approaches for Landsat-7 satellite data

dc.contributor.authorYin, Gaohong
dc.contributor.authorMariethoz, Gregoire
dc.contributor.authorSun, Ying
dc.contributor.authorMcCabe, Matthew
dc.contributor.departmentBiological and Environmental Sciences and Engineering (BESE) Division
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
dc.contributor.departmentEarth System Observation and Modelling
dc.contributor.departmentEnvironmental Science and Engineering
dc.contributor.departmentEnvironmental Science and Engineering Program
dc.contributor.departmentEnvironmental Statistics Group
dc.contributor.departmentStatistics Program
dc.contributor.departmentWater Desalination and Reuse Research Center (WDRC)
dc.contributor.institutionDepartment of Civil and Environmental Engineering, University of Maryland, College Park, MD, USA
dc.contributor.institutionInstitute of Earth Surface Dynamics, University of Lausanne, Lausanne, Switzerland
dc.date.accessioned2020-06-28T09:55:58Z
dc.date.available2020-06-28T09:55:58Z
dc.date.issued2017
dc.description.abstractThe purpose of this study is to assess the relative performance of four different gap-filling approaches across a range of land-surface conditions, including both homogeneous and heterogeneous areas as well as in scenes with abrupt changes in landscape elements. The techniques considered in this study include: (1) Kriging and co-Kriging; (2) geostatistical neighbourhood similar pixel interpolator (GNSPI); (3) a weighted linear regression (WLR) algorithm; and (4) the direct sampling (DS) method. To examine the impact of image availability and the influence of temporal distance on the selection of input training data (i.e. time separating the training data from the gap-filled target image), input images acquired within the same season (temporally close) as well as in different seasons (temporally far) to the target image were examined, as was the case of using information only within the target image itself. Root mean square error (RMSE), mean spectral angle (MSA), and coefficient of determination ($\textit{R}$$^{2}$) were used as the evaluation metrics to assess the prediction results. In addition, the overall accuracy (OA) and kappa coefficient ($\textit{kappa}$) were used to assess a land-cover classification based on the gap-filled images. Results show that all of the gap-filling approaches provide satisfactory results for the homogeneous case, with $\textit{R}$$^{2}$ > 0.93 for bands 1 and 2 in all cases and $\textit{R}$$^{2}$ > 0.80 for bands 3 and 4 in most cases. For the heterogeneous example, GNSPI performs the best, with $\textit{R}$$^{2}$ > 0.85 for all tested cases. WLR and GNSPI exhibit equivalent accuracy when a temporally close input image is used (i.e. WLR and GNSPI both have an $\textit{R}$$^{2}$ equal to 0.89 for band 1). For the case of abrupt changes in scene elements or in the absence of ancillary data, the DS approach outperforms the other tested methods.
dc.identifier.citationGaohong Yin, Gregoire Mariethoz, Sun, Y., & McCabe, M. F. (2017). A comparison of gap-filling approaches for Landsat-7 satellite data. Taylor & Francis. https://doi.org/10.6084/M9.FIGSHARE.5297071.V1
dc.identifier.doi10.6084/m9.figshare.5297071.v1
dc.identifier.urihttp://hdl.handle.net/10754/663884
dc.publisherfigshare
dc.relation.issupplementtoDOI:10.1080/01431161.2017.1363432
dc.subjectSpace Science
dc.subject59999 Environmental Sciences not elsewhere classified
dc.subjectSociology
dc.subject69999 Biological Sciences not elsewhere classified
dc.subjectScience Policy
dc.titleA comparison of gap-filling approaches for Landsat-7 satellite data
dc.typeDataset
display.details.left<span><h5>Type</h5>Dataset<br><br><h5>Authors</h5><a href="https://repository.kaust.edu.sa/search?query=orcid.id:0000-0002-0234-0688&spc.sf=dc.date.issued&spc.sd=DESC">Yin, Gaohong</a> <a href="https://orcid.org/0000-0002-0234-0688" target="_blank"><img src="https://repository.kaust.edu.sa/server/api/core/bitstreams/82a625b4-ed4b-40c8-865a-d6a5225a26a4/content" width="16" height="16"/></a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.author=Mariethoz, Gregoire,equals">Mariethoz, Gregoire</a><br><a href="https://repository.kaust.edu.sa/search?query=orcid.id:0000-0001-6703-4270&spc.sf=dc.date.issued&spc.sd=DESC">Sun, Ying</a> <a href="https://orcid.org/0000-0001-6703-4270" target="_blank"><img src="https://repository.kaust.edu.sa/server/api/core/bitstreams/82a625b4-ed4b-40c8-865a-d6a5225a26a4/content" width="16" height="16"/></a><br><a href="https://repository.kaust.edu.sa/search?query=orcid.id:0000-0002-1279-5272&spc.sf=dc.date.issued&spc.sd=DESC">McCabe, Matthew</a> <a href="https://orcid.org/0000-0002-1279-5272" target="_blank"><img src="https://repository.kaust.edu.sa/server/api/core/bitstreams/82a625b4-ed4b-40c8-865a-d6a5225a26a4/content" width="16" height="16"/></a><br><br><h5>KAUST Department</h5><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Biological and Environmental Sciences and Engineering (BESE) Division,equals">Biological and Environmental Sciences and Engineering (BESE) Division</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division,equals">Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Earth System Observation and Modelling,equals">Earth System Observation and Modelling</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Environmental Science and Engineering,equals">Environmental Science and Engineering</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Environmental Science and Engineering Program,equals">Environmental Science and Engineering Program</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Environmental Statistics Group,equals">Environmental Statistics Group</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Statistics Program,equals">Statistics Program</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Water Desalination and Reuse Research Center (WDRC),equals">Water Desalination and Reuse Research Center (WDRC)</a><br><br><h5>Date</h5>2017</span>
display.details.right<span><h5>Abstract</h5>The purpose of this study is to assess the relative performance of four different gap-filling approaches across a range of land-surface conditions, including both homogeneous and heterogeneous areas as well as in scenes with abrupt changes in landscape elements. The techniques considered in this study include: (1) Kriging and co-Kriging; (2) geostatistical neighbourhood similar pixel interpolator (GNSPI); (3) a weighted linear regression (WLR) algorithm; and (4) the direct sampling (DS) method. To examine the impact of image availability and the influence of temporal distance on the selection of input training data (i.e. time separating the training data from the gap-filled target image), input images acquired within the same season (temporally close) as well as in different seasons (temporally far) to the target image were examined, as was the case of using information only within the target image itself. Root mean square error (RMSE), mean spectral angle (MSA), and coefficient of determination ($\textit{R}$$^{2}$) were used as the evaluation metrics to assess the prediction results. In addition, the overall accuracy (OA) and kappa coefficient ($\textit{kappa}$) were used to assess a land-cover classification based on the gap-filled images. Results show that all of the gap-filling approaches provide satisfactory results for the homogeneous case, with $\textit{R}$$^{2}$ > 0.93 for bands 1 and 2 in all cases and $\textit{R}$$^{2}$ > 0.80 for bands 3 and 4 in most cases. For the heterogeneous example, GNSPI performs the best, with $\textit{R}$$^{2}$ > 0.85 for all tested cases. WLR and GNSPI exhibit equivalent accuracy when a temporally close input image is used (i.e. WLR and GNSPI both have an $\textit{R}$$^{2}$ equal to 0.89 for band 1). For the case of abrupt changes in scene elements or in the absence of ancillary data, the DS approach outperforms the other tested methods.<br><br><h5>Citation</h5>Gaohong Yin, Gregoire Mariethoz, Sun, Y., &amp; McCabe, M. F. (2017). A comparison of gap-filling approaches for Landsat-7 satellite data. Taylor &amp; Francis. https://doi.org/10.6084/M9.FIGSHARE.5297071.V1<br><br><h5>Publisher</h5><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.publisher=figshare,equals">figshare</a><br><br><h5>DOI</h5><a href="https://doi.org/10.6084/m9.figshare.5297071.v1">10.6084/m9.figshare.5297071.v1</a><br><br><h5>Relations</h5><b> Is Supplement To:</b><br/> <ul> <li><i>[Article]</i> <br/> Yin G, Mariethoz G, Sun Y, McCabe MF (2017) A comparison of gap-filling approaches for Landsat-7 satellite data. International Journal of Remote Sensing 38: 6653–6679. Available: http://dx.doi.org/10.1080/01431161.2017.1363432.. DOI: <a href="https://doi.org/10.1080/01431161.2017.1363432" >10.1080/01431161.2017.1363432</a> HANDLE: <a href="http://hdl.handle.net/10754/625993">10754/625993</a></li></ul></span>
display.relations<b> Is Supplement To:</b><br/> <ul> <li><i>[Article]</i> <br/> Yin G, Mariethoz G, Sun Y, McCabe MF (2017) A comparison of gap-filling approaches for Landsat-7 satellite data. International Journal of Remote Sensing 38: 6653–6679. Available: http://dx.doi.org/10.1080/01431161.2017.1363432.. DOI: <a href="https://doi.org/10.1080/01431161.2017.1363432" >10.1080/01431161.2017.1363432</a> HANDLE: <a href="http://hdl.handle.net/10754/625993">10754/625993</a></li></ul>
kaust.personYin, Gaohong
kaust.personSun, Ying
kaust.personMcCabe, Matthew
orcid.id0000-0002-1279-5272
orcid.id0000-0001-6703-4270
orcid.id0000-0002-0234-0688
Files