How to output every pixel coordinates

Microwave Toolbox
#1

Hi
And sorry for this question as there were few answers on the net but none of them worked for me.
Question =>
Is there a way to make esa SNAP output location of every pixels (for ex in a matrix) within a .xml graph to further manipulate data in other platforms ( for ex matlab)
Already tried approach =>
Outputting data in.tiff format which gives corner values automatcally. Then find those corner points ( as terrain correction make the shape not a square) then interpolate. BUT the resulting pixel do not match with the one created from the pin creation segment.
Aim =>
To match data with SMAP coverage after preprocessing completed within matlab. Because I want to also match with subsetted parts of SMAP hence several SMAP data coverages.

==> May be a way to find a coordinate after preprocessing <==
please help.

#2

It is rarely necessary to supply pixel coordinates manually, as the commonly used formats (BEAM DIMAP, NetCDF-CF, BEAM DIMAP) include georeference metadata.

You should mention the data you are using. Some some sensors provide a “satellite view” array where locations are given either as lat and lon arrays or for a “tie-point grid” which then requires interpolation. For mapped data the coordinates can be calculated, but for each program you need to know how coordinates are interpreted. For a (M.N) array, Matlab assumes rectangular pixels but places the reference points at the corners, so requires (M+1),(N+1) coordinates. Other software uses the pixel center as the reference.

#3

Thank you
In order to process I was using geotiff - bigtiff format for sentinel-1 backscatter results.
I didn’t know matlab’s coordinate selection, so it should account for 1 pixel difference.And I think tie-point grids is used, but I am terrain correcting during the preprocessing steps.

#4

After terrain correction I couldn’t find direct latitude - longitude in the metadata so I am thinking here interpolation from the corner coordinates is being used, am I correct. If so what is the suitable way to find corners after the product has been changed to geotiff bigtiff format.
First nonzero pixel along the edges ?
Or can pinning corner cooardinates be applied to graph ?
Or any other way
Can you help guide me please.

#5

Hi, Do you want kml,kmz information of each pixel?

Mustafa

#6

I don’t think geotiff supports tie-point grids. NetCDF-CF tie-point support is available, but it has been a while since I used Matlab so I you should check if support has been added in a recent version of Matlab.
GMTSAR may be useful.

#7

it can also work I guess. Can we make snap prepare those.

#8

Thank you I could try NETCDF format but during processing steps in the terrain correction step tie-point-grid information is being lost do you know why , or a way to work around?

#9

After terrain correction the tie-point are not necessary anymore, because the data is reprojected to a map (you can select during the terrain correction). From this map information the lat/lon values can be calculated.

I don’t know MatLab, but maybe this can help you to access the lat/lon values.

As alternative you can generate lat/lon bands.
Use the Band Maths (Raster menu) and enter as expression LAT (respectively LON) and give the resulting band a name, ‘latitude’ respectively ‘longitude’.
Now you have two additional bands in your product which provide the geo-location.

#10

Thank you so much. Creating another band is really what I want. For the page there seems to be some problem. Either way thank you so much.

#11

it could be