since I passed to OTB 8, I am experiencing unusually long computation times when doing Orthorectification of Sentinel-1 images using both DEM and a geoid file (egm96.grd with the header file trick mentioned in the OTB doc). I did not measure exactly but we’re talking about a 50x to 100x longer time. Computation time returns reasonable (comparable to OTB 7.4) when I remove the elev.geoid parameter.
Is this a known issue? Fyi, I’m using the orthofit mode (outputs.ortho is a fully overlapping Sentinel-2 scene), and I have no significant differences in the output when using geoid file or not.
Thanks in advance,