Monitoring Land Cover Changes in Jambi Province, Sumatra using Sentinel-1 and Google Earth Engine 15 th of November 2016 Kira Urban and Lutz Fehrmann Email: [email protected] | [email protected]Chair of Forest Inventory and Remote Sensing Faculty of Forest Science and Forest Ecology Georg-August-University Goettingen
22
Embed
Monitoring Land Cover Changes in Jambi Province, Sumatra using Sentinel ...wiki.awf.forst.uni-goettingen.de/wiki/images/2/2a/Uban&fehrmann.pdf · Monitoring Land Cover Changes in
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Monitoring Land Cover Changes in Jambi Province, Sumatra usingSentinel-1 and Google Earth Engine
· Near real-time detection· Land cover change detection→ Burned area· Area calculation· Evaluation of suitability of Sentinel-1
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
Methodology4Study Area
Sentinel-1
Google Earth Engine
Analysis
Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
MethodologyStudy Area
5Sentinel-1
Google Earth Engine
Analysis
Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Hosted by the European Space Ageny (ESA)· C-SAR· Two-satellite constellation· Spatial resolution: up to 5 x 5 m· Temporal resolution: 6 days
(up to 3 days over Europe and Canada)
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
MethodologyStudy Area
Sentinel-1
6Google Earth Engine
Analysis
Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Data download not necessary· Huge computational power· Platform for the scientific analysis of high-resolution imagery· Team of > 100 programmers· Java Script or Python API
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
MethodologyStudy Area
Sentinel-1
6Google Earth Engine
Analysis
Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
MethodologyStudy Area
Sentinel-1
Google Earth Engine
7Analysis
Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Clear cut or burned area of a minimum size is required
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
MethodologyStudy Area
Sentinel-1
Google Earth Engine
7Analysis
Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Clear cut or burned area of a minimum size is required· Sentinel-1 Ground Range Detection (GRD) of 2015
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
MethodologyStudy Area
Sentinel-1
Google Earth Engine
7Analysis
Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Clear cut or burned area of a minimum size is required· Sentinel-1 Ground Range Detection (GRD) of 2015· Filtering for Interferometric Wide Swath Mode (IW), dual polarization (VH)
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
MethodologyStudy Area
Sentinel-1
Google Earth Engine
7Analysis
Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Clear cut or burned area of a minimum size is required· Sentinel-1 Ground Range Detection (GRD) of 2015· Filtering for Interferometric Width (IW), dual polarization (VH)· Processing monthly mean SAR imagery· Calculating the differences between two months
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
MethodologyStudy Area
Sentinel-1
Google Earth Engine
7Analysis
Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Clear cut or burned area of a minimum size is required· Sentinel-1 Ground Range Detection (GRD) of 2015· Filtering for Interferometric Width (IW), dual polarization (VH)· Processing monthly mean SAR imagery· Calculating the differences between two months· Visual interpretation· Dual temporal composite· Validation: MODIS
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
Methodology
8Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
Methodology
9Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
July - November 2015 August October
September November
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
Methodology
10Results
Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
Methodology
Results
11Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Changes in backscatter are caused by many different aspects:· Acquisition mode· Local topographic aspects like slope or surface roughness· Water content
=⇒ complicate the distinction between actual change in land cover caused byfire or anthropogenic disturbances
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
Methodology
Results
11Discussion
Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Changes in backscatter are caused by many different aspects:· Acquisition mode· Local topographic aspects like slope or surface roughness· Water content
=⇒ complicate the distinction between actual change in land cover caused byfire or anthropogenic disturbances· Validation data
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
Methodology
Results
Discussion
12Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Earth Engine provides preprocessed data (Level 1 GRD)· opportunity for a fast processing· disable analysis based on Single Look Complex (SLC) data
· Powerful tool for change detection with a high spatial and temporal resolution
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
Methodology
Results
Discussion
12Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Earth Engine provides preprocessed data (Level 1 GRD)· opportunity for a fast processing· disable analysis based on Single Look Complex (SLC) data
· Powerful tool for change detection with a high spatial and temporal resolution· This study showed successfully the suitability of Sentinel-1 data for monitoring
burned areas and its expansion on a high temporal resolution· Radar imagery is a beneficial tool for observing land cover change
12
Monitoring Land CoverChanges using
Sentinel-1 and GoogleEarth Engine
Kira Urban andLutz Fehrmann
Background
Objectives
Methodology
Results
Discussion
12Conclusion andOutlook
Chair of Forest Inventoryand Remote Sensing
Georg-August-UniversityGoettingen
· Earth Engine provides preprocessed data (Level 1 GRD)· opportunity for a fast processing· disable analysis based on Single Look Complex (SLC) data
· Powerful tool for change detection with a high spatial and temporal resolution· This study showed successfully the suitability of Sentinel-1 data for monitoring
burned areas and its expansion on a high temporal resolution· Radar imagery is a beneficial tool for observing land cover change