Skip to main content
  • EN

I'm trying to use RasterClipOverlay to extract an area on GeoTIFF defined by a polygon.  FME is currently giving me the incorrect area.  The Workspace:

 

The workspace has a polygon coming from a POSTGIS layer which is filtered. It also has an GeoTIFF  both of which are inputs to the RasterClipOverlay.  The results are output to a GEOTIFF.  As a test I have also added a simple output image that does not come from the RasterClipOverlay. Both the polygon and image are in the same projection EPSG:3395.

The original image (displayed in QGIS):  

 

The polygon (in QGIS):

The results: 
Displayed in QGIS with Open Street Map behind and the polygon overlaid.
​​​​As you see from the results, the area of the map is not correctly extracted.  The polygon is supposed to cut out the sea, however, it also cuts the map boundaries and part of the land. The shape of the cut is correct however the map is “shrunk” hence the area of the map is incorrect. 

This is just one example, I have various other maps I’m working with and the problem occurs in all of them.  If I use QGIS /GDAL then the results are correct, however, I want this to work in FME.

The test output that does not get clipped does display the correct map in the correct location.

 

Can you check both  the Geotiff and the PostGIS raster are in the same coordinate system? Maybe also desegregate all the polygons, if there are multi part polygons

I have checked the coordinate reference system and both are in the same 3395.  I had included transformers to ensure this but they didn’t make any difference so removed them to keep my model simple as possible.

 

I have since ensure my polygons and not multi and that has not made any difference either.

Out of interest, how are you checking the coordinate system? You should inspect both inputs into the problem transformer, opening the ‘Feature Information’. Please send screenshots of the geometry section of the feature information window

Out of interest, how are you checking the coordinate system? You should inspect both inputs into the problem transformer, opening the ‘Feature Information’. Please send screenshots of the geometry section of the feature information window

 

Out of interest, how are you checking the coordinate system? You should inspect both inputs into the problem transformer, opening the ‘Feature Information’. Please send screenshots of the geometry section of the feature information window

 

I have updated the workspace with Coordinate setters to be sure the CRS is used.  However, the end result is the same.  I can confirm that the data is the CRS that I applied with these transformers.  Running SELECT DISTINCT(st_srid(geom) FROM TABLE_NAME on the database returns 3395.  If I inspect the image directly before any transformer it also in WORLD MERCATOR 3395.

  I emphasise the GDAL is correctly handling the same task.

Is there a reason why you use the FME Hub transformer RasterClipOverlay instead of the official Clipper transformer?

Is there a reason why you use the FME Hub transformer RasterClipOverlay instead of the official Clipper transformer?

Thank you, this is probably the solution.  I had forgotten I could use the clipper for raster.   However, how can I use Clipper and ensure the outside area of the image is transparent?  This is something that RasterClipOverlay would do automatically.

According to the documentation: "Add Alpha/Nodata as Yes will add a transparency layer where cells fall outside the irregular clip boundary but inside the raster extents”.

Reply


Community Stats

31,821
Posts
120,861
Replies
39,452
Members

Latest FME

Terms & ConditionsCookie settings
A product of
Safe Software respectfully acknowledges that we live, learn and work on the traditional and unceded territories of the Kwantlen, Katzie, and Semiahmoo First Nations.