Warping Radarsat - 2 Images

Hi Everyone,

I’m working with some radarsat - 2 data, and everything seemed to be good with my snap process until I switched the calibration to output in dBs.

Now, when I bring it into ENVI or arcmap, the image seems warped at the upper left hand corner. Please see attached.

My basic process is:
Radiometric calibration > output in dBs
speckle filter > refined lee
multilook > 4x4
geometric calibration> elipsoid correction> geolocation grid> bilinear interpolation, UTM/WGS84(automatic)> saving as a geotiff.

I’m still new to SNAP and any help would be much appreciated.


Can’t figure which area I see. The distortion does not look strange to me if the area is located in the higher latitudes.

Conversion to dB should be done in the end.

Ah okay. So does that mean then there is a tool in snap to perform the same workflow then output to dB in the end? Or should I be doing my radiometric calibration last?

Thanks for the help!

Agreed - and it is at higher lat. The issue that I think is strange is that when not outputting to dB in my radiometric calibration, I dont have such an issue with this datum and projection.

Calibration should be done early and dB-conversion as late as possible. Many spatial operations (for example simple average) give incorrect results if done in dB scale. Ideally the toolbox should check (and autoconvert internally) to and from dB (there are operators for this under Raster/Data Conversion) but I guess it’s possible that some operators do not check for this.

Yes, that makes good sense.

Though I’m having trouble using the operator under raster/data conversion on my post geometric calibration raster. The linear to dB operator is greyed out and I cant seem to make it available regardless of file type.

Any suggestions? I’ve been outputting to a tiff

You should use the internal format (BEAM-Dimap) inside the toolbox and convert to tiff when you export.

I appreciate your help.

Strange, the linear to dB operator is still greyed out, and deriving a histogram from the image confirms that it is not in dB.