SAR change detection with SNAP

From AWF-Wiki
(Difference between revisions)
Jump to: navigation, search
(Data transfer)
(Display in QGIS)
 
(27 intermediate revisions by one user not shown)
Line 1: Line 1:
In this tutorial we will use Level-1 Ground Range Detected (GRD) Sentinel-1 data downloaded from the Copernicus Open Access Hub. The example are fire events during 2015 in a peat forest located in Eastern Jambi, Sumatra, Indonesia.
+
In this tutorial we will use Level-1 Ground Range Detected (GRD) Sentinel-1 data downloaded from the Copernicus Open Access Hub. The example are fire events during 2019 in a peat forest located in Eastern Jambi, Sumatra, Indonesia.
  
 
=Data transfer=
 
=Data transfer=
Download [http://www.gwdg.de/~hfuchs/qgis/geodata_jambi.zip example tutorial data from Jambi] or data from your personal region of interest: two Sentinel-1 scenes one pre-event acquisition date and one post-event date. Follow and adapt accordingly the instruction on [[Downloading Sentinel-2 images]].
+
Download example data from STud.IP or use data from your personal region of interest: two Sentinel-1 scenes one pre-event acquisition date and one post-event date. Follow and adapt accordingly the instruction on [[Downloading Sentinel-2 images]].
 
Sentinel-1 data can have different characteristics (product type, polarization, sensor mode and ascending/descending orbit). Make sure, that you have the same characteristics for both images.
 
Sentinel-1 data can have different characteristics (product type, polarization, sensor mode and ascending/descending orbit). Make sure, that you have the same characteristics for both images.
 
Set the sensing period (e.g. from 01.06.2019 to 30.11.2019).
 
Set the sensing period (e.g. from 01.06.2019 to 30.11.2019).
Line 13: Line 13:
 
* Open SNAP Desktop.  
 
* Open SNAP Desktop.  
 
* Go to {{mitem|text=File → Import → SAR sensors → Sentinel-1}} and select the downloaded zip files.
 
* Go to {{mitem|text=File → Import → SAR sensors → Sentinel-1}} and select the downloaded zip files.
* Unfold the product folders and Mark the Aplitude_VV bands of both products.
+
* In '''Product Explorer''' click on + to unfold the product folders. Mark the '''Amplitude_VV''' bands of both products.
 
* Right click '''Open 2 Image Window'''.
 
* Right click '''Open 2 Image Window'''.
* {{mitem|text=Windows → Tile Horizontally}}. Both Amplitude bands are shown side by side geometrically linked to each other. The geometry ist still in sensor coordinates in Ascending mode.  
+
* {{mitem|text=Windows → Tile Horizontally}}. Both Amplitude bands are shown side by side geometrically linked to each other. Make sure the option {{mitem|text=View --> Synchronise Image View}} and  {{mitem|text=View --> Synchronise Image Cursors}}is activated).
 +
* Adjust the viewer size that they are evenly.
 +
* Zoom to  an area of interest where you may detect changes (see below). Compare the pre-event (2019-06-06) and post-event image (2019-11-21).  
 
[[File:snap_sidebyside.png|800px]]
 
[[File:snap_sidebyside.png|800px]]
  
 
=Subsetting=  
 
=Subsetting=  
To reduce the data volume subset the  Sentinel-1 scenes.
+
Subset the  Sentinel-1 scenes reducing the data volume.
* Close horizontal tileing.
+
* Activate the pre-event image clicking on the tab of the viewer. The active image is marked with a yellow frame.
* Zoom to your area of interest.
+
* Right click into the image and select '''Spatial Subset from View...'''.
+
  
 
[[File:SNAP_Select_Subset_Area.PNG|600px]]
 
[[File:SNAP_Select_Subset_Area.PNG|600px]]
  
* Besides a spatial subset you could also select for single bands with this tool. Go to {{button|text=Bands subset}} and select '''Amplitude_VV'''.
+
* Right click into the active image and select '''Spatial Subset from View...'''.
* Click {{button|text=OK}}.
+
  
 
[[File:SNAP_Subset.PNG|400px]]
 
[[File:SNAP_Subset.PNG|400px]]
  
 
+
* Besides a spatial subset you can also select a band subset with this tool. Go to the tab {{button|text=Band subset}} and select only one band '''Amplitude_VV'''.
 
+
* Click {{button|text=OK}}.
'''Please note: The image can still be in acquisition geometry, whereby the image appears mirrored.'''
+
* A new Product subset_* can be found in the Product Explorer.
 +
+ Activate the prost-event image and repeat the work steps above.
 +
''Note, the resulting images are still in acquisition geometry, whereby the images appear mirrored.''
  
 
=Radiometric and geometric calibration=
 
=Radiometric and geometric calibration=
 
=== Apply Orbit File ===  
 
=== Apply Orbit File ===  
* Select the subsetted product from the previous step within the '''Product Explorer''' on the left side.
+
* Select the subset '''pre-event''' product from the previous step within the '''Product Explorer'''.  
* Go to {{mitem|text=Radar → Apply Orbit File}}
+
* Go to {{mitem|text=Radar → Apply Orbit File}}  
 
* Default settings are fine.
 
* Default settings are fine.
 
* Select target product directory.
 
* Select target product directory.
Line 46: Line 47:
 
=== Radiometric Calibration ===  
 
=== Radiometric Calibration ===  
 
* Select the product from the previous step within the '''Product Explorer''' on the left side.
 
* Select the product from the previous step within the '''Product Explorer''' on the left side.
* Go to {{mitem|text=Radar → Radiometric Calibration → Calibrate}}
+
* Go to {{mitem|text=Radar → Radiometric → Calibrate}}
 
* Select target product directory.
 
* Select target product directory.
 
* Go to the '''Processing Parameters''' tab and check sigma0. (You could also use gamma0 processing.)
 
* Go to the '''Processing Parameters''' tab and check sigma0. (You could also use gamma0 processing.)
Line 55: Line 56:
 
* Select the product from the previous step within the '''Product Explorer''' on the left side.
 
* Select the product from the previous step within the '''Product Explorer''' on the left side.
 
* For speckle filtering, there are many different options of filter algorithms and window size.
 
* For speckle filtering, there are many different options of filter algorithms and window size.
* Use the Lee filter and try a window size of 3x3 and 9x9, what are the differences in your result?
+
* Use the Lee filter with a window size of 3x3.
 
* Go to {{mitem|text=Radar → Speckle Filtering → Single Product Speckle Filter}}
 
* Go to {{mitem|text=Radar → Speckle Filtering → Single Product Speckle Filter}}
 
* Select target product directory.
 
* Select target product directory.
Line 72: Line 73:
 
* Now the image should be correctly displayed, in terms of geometry, and be smoothed.
 
* Now the image should be correctly displayed, in terms of geometry, and be smoothed.
  
Repeat all steps for the second acquisiton date product!
+
'''Repeat the radiometric and geometric calibration steps for the post-event subset (second acquisiton date)!'''
  
 
=Change composite=
 
=Change composite=
 
=== Create Stack ===
 
=== Create Stack ===
 
* Close all products by marking them all '''Product Explorer''' on the left side, right click and '''Close All Products'''.
 
* Close all products by marking them all '''Product Explorer''' on the left side, right click and '''Close All Products'''.
* Go to {{mitem|text=File → Open}} and select both radiometric and geometric corrected images from the previous step with the extension '''_Orb_Cal_Spk_TC.dim'''.
+
* {{mitem|text=File --> Open Product}}. Load both radiometric and geometric corrected images from the previous step with the extension '''_Orb_Cal_Spk_TC.dim'''.
 
* Go to {{mitem|text=Radar → Coregistration → Stack Tools → Create Stack}}
 
* Go to {{mitem|text=Radar → Coregistration → Stack Tools → Create Stack}}
* Click on {{button|text=Add Opened}} button on the right side to select the images.
+
* Click on [[File:snap_add_open_button.png]] button on the right side to select the images.
 
* Leave the default settings at the '''2-CreateStack''' tab.
 
* Leave the default settings at the '''2-CreateStack''' tab.
 
* Go to the '''3-Write''' tab and select target product directory.
 
* Go to the '''3-Write''' tab and select target product directory.
Line 87: Line 88:
 
=== Convert Datatype ===
 
=== Convert Datatype ===
 
* Select the product from the previous step within the '''Product Explorer''' on the left side.
 
* Select the product from the previous step within the '''Product Explorer''' on the left side.
* Go to {{mitem|text=Raster → Geometric → Data Conversion → Convert datatype}}
+
* Go to {{mitem|text=Raster → Data Conversion → Convert datatype}}
 
* Save as {{button|text=GeoTIFF}}.
 
* Save as {{button|text=GeoTIFF}}.
 
* Select target product directory.
 
* Select target product directory.
Line 98: Line 99:
 
* {{mitem|text=Data Source Manger --> Browser --> XYZ Tiles --> Google Satellite}}
 
* {{mitem|text=Data Source Manger --> Browser --> XYZ Tiles --> Google Satellite}}
 
* Load the change compoisite file  with extension '''_Cnv.tif''' {{mitem|text=Data Source Manger --> Raster}}.
 
* Load the change compoisite file  with extension '''_Cnv.tif''' {{mitem|text=Data Source Manger --> Raster}}.
* {{mitem|text=Layer properties --> Symbology}}. Adjust the band assignment: Red = Band 2, Green = Band 1, Blue = Band 1.  
+
* {{mitem|text=Layer properties --> Symbology}}. Adjust the band assignment: Red = Band 2, Green = Band 1, Blue = Band 1.
 
* Adjust opacity of the change layer on top of Google Satellite. {{mitem|text=Layer Properties --> Transparency --> Global Transparency}}.
 
* Adjust opacity of the change layer on top of Google Satellite. {{mitem|text=Layer Properties --> Transparency --> Global Transparency}}.
 
In the bi-temporal change composite red indicates change areas.
 
In the bi-temporal change composite red indicates change areas.

Latest revision as of 14:16, 25 January 2021

In this tutorial we will use Level-1 Ground Range Detected (GRD) Sentinel-1 data downloaded from the Copernicus Open Access Hub. The example are fire events during 2019 in a peat forest located in Eastern Jambi, Sumatra, Indonesia.

Contents

[edit] Data transfer

Download example data from STud.IP or use data from your personal region of interest: two Sentinel-1 scenes one pre-event acquisition date and one post-event date. Follow and adapt accordingly the instruction on Downloading Sentinel-2 images. Sentinel-1 data can have different characteristics (product type, polarization, sensor mode and ascending/descending orbit). Make sure, that you have the same characteristics for both images. Set the sensing period (e.g. from 01.06.2019 to 30.11.2019). Select Product type GRD and sensor mode IW. Download the images of the 06.06.2019 and 21.11.2019 (both GRD, VV/VH, IW, Descending).

Esa openhub S1.png

[edit] Visualization of amplitude bands

  • Open SNAP Desktop.
  • Go to File → Import → SAR sensors → Sentinel-1 and select the downloaded zip files.
  • In Product Explorer click on + to unfold the product folders. Mark the Amplitude_VV bands of both products.
  • Right click Open 2 Image Window.
  • Windows → Tile Horizontally. Both Amplitude bands are shown side by side geometrically linked to each other. Make sure the option View --> Synchronise Image View and View --> Synchronise Image Cursorsis activated).
  • Adjust the viewer size that they are evenly.
  • Zoom to an area of interest where you may detect changes (see below). Compare the pre-event (2019-06-06) and post-event image (2019-11-21).

Snap sidebyside.png

[edit] Subsetting

Subset the Sentinel-1 scenes reducing the data volume.

  • Activate the pre-event image clicking on the tab of the viewer. The active image is marked with a yellow frame.

SNAP Select Subset Area.PNG

  • Right click into the active image and select Spatial Subset from View....

SNAP Subset.PNG

  • Besides a spatial subset you can also select a band subset with this tool. Go to the tab Band subset and select only one band Amplitude_VV.
  • Click OK.
  • A new Product subset_* can be found in the Product Explorer.

+ Activate the prost-event image and repeat the work steps above. Note, the resulting images are still in acquisition geometry, whereby the images appear mirrored.

[edit] Radiometric and geometric calibration

[edit] Apply Orbit File

  • Select the subset pre-event product from the previous step within the Product Explorer.
  • Go to Radar → Apply Orbit File
  • Default settings are fine.
  • Select target product directory.

Snap apply orbit.png

  • Click Run.

[edit] Radiometric Calibration

  • Select the product from the previous step within the Product Explorer on the left side.
  • Go to Radar → Radiometric → Calibrate
  • Select target product directory.
  • Go to the Processing Parameters tab and check sigma0. (You could also use gamma0 processing.)

Snap calibrate.png

  • Click Run.

[edit] Single Product Speckle Filter

  • Select the product from the previous step within the Product Explorer on the left side.
  • For speckle filtering, there are many different options of filter algorithms and window size.
  • Use the Lee filter with a window size of 3x3.
  • Go to Radar → Speckle Filtering → Single Product Speckle Filter
  • Select target product directory.
  • Go to the Processing Parameters tab and select Lee in the Filter drop down menue.
  • Adapt the window size in X and Y direction accordingly.

Snap leefilter.png

  • Click Run.

[edit] Range-Doppler-Terrain Correction

  • Select the product from the previous step within the Product Explorer on the left side.
  • Go to Radar → Geometric → Terrain Correction → Range-Doppler-Terrain Correction
  • Select target product directory.
  • Go to the Processing Parameters tab and select SRTM 1sec HGT (Auto Download) in the Digital Elevation Model drop down menue.

Snap terrain correction.png

  • Click Run.
  • Now the image should be correctly displayed, in terms of geometry, and be smoothed.

Repeat the radiometric and geometric calibration steps for the post-event subset (second acquisiton date)!

[edit] Change composite

[edit] Create Stack

  • Close all products by marking them all Product Explorer on the left side, right click and Close All Products.
  • File --> Open Product. Load both radiometric and geometric corrected images from the previous step with the extension _Orb_Cal_Spk_TC.dim.
  • Go to Radar → Coregistration → Stack Tools → Create Stack
  • Click on Snap add open button.png button on the right side to select the images.
  • Leave the default settings at the 2-CreateStack tab.
  • Go to the 3-Write tab and select target product directory.

Snap stack.png

  • Click Run.

[edit] Convert Datatype

  • Select the product from the previous step within the Product Explorer on the left side.
  • Go to Raster → Data Conversion → Convert datatype
  • Save as GeoTIFF.
  • Select target product directory.

Snap datatype2.png

  • Default settings are fine.
  • Click Run.

[edit] Display in QGIS

  • Open QGIS.
  • Data Source Manger --> Browser --> XYZ Tiles --> Google Satellite
  • Load the change compoisite file with extension _Cnv.tif Data Source Manger --> Raster.
  • Layer properties --> Symbology. Adjust the band assignment: Red = Band 2, Green = Band 1, Blue = Band 1.
  • Adjust opacity of the change layer on top of Google Satellite. Layer Properties --> Transparency --> Global Transparency.

In the bi-temporal change composite red indicates change areas.

Snap change composite2.png

Personal tools
Namespaces

Variants
Actions
Navigation
Development
Toolbox
Print/export