[gdal-dev] OGRCreateCoordinateTransformation()
Steve Riddell
sriddell at geocue.com
Mon Mar 28 11:37:08 PDT 2022
Hi Even,
Sorry for the delay, but I’ve now tested with PR3216, alone, and then with PR3127. I’m getting the same behavior as I reported below. Perhaps my best chance at getting this to work is to either directly make proj calls, or construct the pipeline myself and pass through gdal?
Steve
From: Even Rouault <even.rouault at spatialys.com>
Sent: Thursday, March 24, 2022 10:50 AM
To: Steve Riddell <sriddell at geocue.com>; proj <proj at lists.osgeo.org>
Cc: gdal-dev at lists.osgeo.org
Subject: Re: OGRCreateCoordinateTransformation()
Steve,
(adding proj mailing list since it's actually purely a PROJ topic
Does the fact that there isn’t a consensus mean 3126 & 3127 aren’t going to end up in a PROJ release?
Did you try them to confirm they help ? Maybe if you want to weight in the PRs (if they actually help) that could change the balance.
I read through the PR’s. Are you thinking my issue is that a vertical shift is being introduced because of a transformation to WGS84 (from NAD83(CSRS)) to look up the geoid separation value?
That's probably an issue with the wrong interpolation CRS being used, but I can't generally predict how this complex code works without launching a debugger and step-by-step tracing what happens.
PROJ4_GRIDS provided a way to specify exactly which geoid file(s) to use. Some vertical datums, such as NAVD88, only have a single EPSG code for multiple geoids – what’s the replacement for PROJ4_GRIDS in PROJ9? If I could switch to WKT2 strings (not sure that’s really feasible) does it fix these problems?
WKT2 hardly helps for that. The closest equivalent is the GEOIDMODEL[] sub node of a VERTCRS[] (http://docs.opengeospatial.org/is/18-010r7/18-010r7.html#83). But that's something that refer to an existing transformation . And in a VRTCRS[] you can actually have several GEOIDMODEL[] listed (e.g. for NAVD88, all the geoid models), since that's more a definition of the VertCRS than which particular geoid model to use. So that's not nominally appropriate for a "random" grid not known of PROJ, although we have a bit bent the concept to allow a GEOIDMODEL["PROJ name_of_the_grid_dot_tif_or_gtx"]. But I would suspect that in that situation the ambiguity with the geographic / interpolation CRS to use would be similar to WKT1 PROJ4_GRIDS.
Thanks,
steve
From: Even Rouault <even.rouault at spatialys.com><mailto:even.rouault at spatialys.com>
Sent: Wednesday, March 23, 2022 6:06 PM
To: Steve Riddell <sriddell at geocue.com><mailto:sriddell at geocue.com>; gdal-dev at lists.osgeo.org<mailto:gdal-dev at lists.osgeo.org>
Subject: Re: OGRCreateCoordinateTransformation()
Steve,
you may try PR https://github.com/OSGeo/PROJ/pull/3126 or https://github.com/OSGeo/PROJ/pull/3127, that might potentially help, but they don't make consensus (see discussions of those PRs). Basically it is a nightmare to deal with the old TOWGS84 / PROJ4_GRIDS approach, that poorly specifies some elements of transformations (especially the vertical part), and doesn't fit well in the ISO-19111 model used by newer PROJ (we use and abuse the BoundCRS approach in complex ways). One of the issue too is that we've realized late in the process that we didn't emulate some aspects of the old transformation pipeline, and there isn't general enthusiasm to change that now.
Even
Le 23/03/2022 à 23:42, Steve Riddell a écrit :
Hi Even,
I tried the fix with the CRS’s I sent, and it worked. But as I mentioned, my real target CRS is a custom projected system (below), and it still fails with what seems to be the same error as in my original email.
COMPD_CS["TempTM + CGVD28 height - HT2_0",
PROJCS["Custom",
GEOGCS["NAD83(CSRS)",
DATUM["NAD83_Canadian_Spatial_Reference_System",
SPHEROID["GRS 1980",6378137,298.257222101,
AUTHORITY["EPSG","7019"]],
TOWGS84[0,0,0,0,0,0,0],
AUTHORITY["EPSG","6140"]],
PRIMEM["Greenwich",0,
AUTHORITY["EPSG","8901"]],
UNIT["degree",0.0174532925199433,
AUTHORITY["EPSG","9122"]],
AUTHORITY["EPSG","4617"]],
PROJECTION["Transverse_Mercator"],
PARAMETER["latitude_of_origin",49.351346659616],
PARAMETER["central_meridian",-123.20266499149],
PARAMETER["scale_factor",1],
PARAMETER["false_easting",15307.188],
PARAMETER["false_northing",6540.975],
UNIT["Meters",1],
AXIS["Easting",EAST],
AXIS["Northing",NORTH]],
VERT_CS["CGVD28 height - HT2_0",
VERT_DATUM["Canadian Geodetic Vertical Datum of 1928",2005,
EXTENSION["PROJ4_GRIDS","HT2_0.gtx"],
AUTHORITY["EPSG","5114"]],
UNIT["metre",1,
AUTHORITY["EPSG","9001"]],
AXIS["Gravity-related height",UP],
AUTHORITY["EPSG","5713"]]]
Experimenting, I removed the “TOWGS84” directive from BOTH the source and target systems, and I do get an answer instead of an error:
For test point: (49.351346658, -123.202664992, 289.809)
I get: (15307.1880,6540.9748,307.744)
But since geoid separation at the test point is -18.238, Z should be 308.047. In fact, if I code up the following pipeline and initialize a transform object:
strPipe = "+proj=pipeline\
+step +proj=axisswap +order=2, 1\
+step +proj=unitconvert +xy_in=deg +xy_out=rad\
+step +inv +proj=vgridshift +grids=HT2_0.gtx +multiplier=1\
+step +proj=tmerc +lat_0=49.351346659616 +lon_0=-123.20266499149 +k=1\
+x_0=15307.188 +y_0=6540.975 +ellps=GRS80";
The call to Transform() gives: (15307.1880, 6540.9748, 308.042) -- XY matches above, but now I get the expected Z (-0.005). However, I’d like to avoid assembling pipelines for every conversion/transform.
Any arrows left in your quiver?
Thanks,
Steve
From: Steve Riddell
Sent: Friday, March 18, 2022 6:17 PM
To: Even Rouault <even.rouault at spatialys.com><mailto:even.rouault at spatialys.com>; gdal-dev at lists.osgeo.org<mailto:gdal-dev at lists.osgeo.org>
Subject: RE: OGRCreateCoordinateTransformation()
Thanks very much, Even. For reporting the problem, I “packaged up” a case with CRS’s with EPGS codes. For the actual problem I’m working, the target CRS is a custom projection –no EPSG.
From: Even Rouault <even.rouault at spatialys.com<mailto:even.rouault at spatialys.com>>
Sent: Friday, March 18, 2022 5:05 PM
To: Steve Riddell <sriddell at geocue.com<mailto:sriddell at geocue.com>>; gdal-dev at lists.osgeo.org<mailto:gdal-dev at lists.osgeo.org>
Subject: Re: OGRCreateCoordinateTransformation()
Steve,
Fix in PROJ queued in https://github.com/OSGeo/PROJ/pull/3123
The complexity of dealing with legacy features TOWGS84[] and PROJ4_GRIDS is highly stressing PROJ pipeline computation engine.
Using the plain EPSG codes EPSG:4955 -> EPSG:4617+5713 would be much preferable here
Even
Le 18/03/2022 à 21:14, Steve Riddell a écrit :
Hi,
I’m getting an error (below) using GDAL trying to set up a transformation between NAD83(CSRS) geodetic, w/ ellipsoidal heights, and NAD83(CSRS) geodetic, CGVD28 heights. Similar transformations work fine on some datums, but fail on others. Hope someone can offer some insight. Note the if I don't use the PROJ4_GRIDS extension, I don't get the error, but I don't get the geoid adjustment either.
COMPD_CS["NAD83(CSRS) + Ellipsoid (Meters)",
GEOGCS["NAD83(CSRS)",
DATUM["NAD83_Canadian_Spatial_Reference_System",
SPHEROID["GRS 1980",6378137,298.257222101,
AUTHORITY["EPSG","7019"]],
TOWGS84[0,0,0,0,0,0,0],
AUTHORITY["EPSG","6140"]],
PRIMEM["Greenwich",0,
AUTHORITY["EPSG","8901"]],
UNIT["degree",0.0174532925199433,
AUTHORITY["EPSG","9122"]],
AUTHORITY["EPSG","4617"]],
VERT_CS["Ellipsoid (Meters)",
VERT_DATUM["Ellipsoid",2002],
UNIT["metre",1,
AUTHORITY["EPSG","9001"]],
AXIS["ellipsoidal height",UP]]]
COMPD_CS["NAD83(CSRS) + CGVD28 height - HT2_0",
GEOGCS["NAD83(CSRS)",
DATUM["NAD83_Canadian_Spatial_Reference_System",
SPHEROID["GRS 1980",6378137,298.257222101,
AUTHORITY["EPSG","7019"]],
TOWGS84[0,0,0,0,0,0,0],
AUTHORITY["EPSG","6140"]],
PRIMEM["Greenwich",0,
AUTHORITY["EPSG","8901"]],
UNIT["degree",0.0174532925199433,
AUTHORITY["EPSG","9122"]],
AUTHORITY["EPSG","4617"]],
VERT_CS["CGVD28 height - HT2_0",
VERT_DATUM["Canadian Geodetic Vertical Datum of 1928",2005,
EXTENSION["PROJ4_GRIDS","HT2_0.gtx"],
AUTHORITY["EPSG","5114"]],
UNIT["metre",1,
AUTHORITY["EPSG","9001"]],
AXIS["Gravity-related height",UP],
AUTHORITY["EPSG","5713"]]]
ERROR 6: Cannot find coordinate operations from `EPSG:4955' to `COMPOUNDCRS["NAD83(CSRS) + CGVD28 height - HT2_0",BOUNDCRS[SOURCECRS[GEOGCRS["NAD83(CSRS)",DATUM["NAD83 Canadian Spatial Reference System",ELLIPSOID["GRS 1980",6378137,298.257222101,LENGTHUNIT["metre",1]]],PRIMEM["Greenwich",0,ANGLEUNIT["degree",0.0174532925199433]],CS[ellipsoidal,2],AXIS["geodetic latitude (Lat)",north,ORDER[1],ANGLEUNIT["degree",0.0174532925199433]],AXIS["geodetic longitude (Lon)",east,ORDER[2],ANGLEUNIT["degree",0.0174532925199433]],ID["EPSG",4617]]],TARGETCRS[GEOGCRS["WGS 84",DATUM["World Geodetic System 1984",ELLIPSOID["WGS 84",6378137,298.257223563,LENGTHUNIT["metre",1]]],PRIMEM["Greenwich",0,ANGLEUNIT["degree",0.0174532925199433]],CS[ellipsoidal,2],AXIS["latitude",north,ORDER[1],ANGLEUNIT["degree",0.0174532925199433]],AXIS["longitude",east,ORDER[2],ANGLEUNIT["degree",0.0174532925199433]],ID["EPSG",4326]]],ABRIDGEDTRANSFORMATION["Transformation from NAD83(CSRS) to WGS84",METHOD["Position Vector transformation (geog2D domain)",ID["EPSG",9606]],PARAMETER["X-axis translation",0,ID["EPSG",8605]],PARAMETER["Y-axis translation",0,ID["EPSG",8606]],PARAMETER["Z-axis translation",0,ID["EPSG",8607]],PARAMETER["X-axis rotation",0,ID["EPSG",8608]],PARAMETER["Y-axis rotation",0,ID["EPSG",8609]],PARAMETER["Z-axis rotation",0,ID["EPSG",8610]],PARAMETER["Scale difference",1,ID["EPSG",8611]]]],BOUNDCRS[SOURCECRS[VERTCRS["CGVD28 height - HT2_0",VDATUM["Canadian Geodetic Vertical Datum of 1928"],CS[vertical,1],AXIS["gravity-related height",up,LENGTHUNIT["metre",1]],ID["EPSG",5713]]],TARGETCRS[GEOGCRS["WGS 84",DATUM["World Geodetic System 1984",ELLIPSOID["WGS 84",6378137,298.257223563,LENGTHUNIT["metre",1]]],PRIMEM["Greenwich",0,ANGLEUNIT["degree",0.0174532925199433]],CS[ellipsoidal,3],AXIS["latitude",north,ORDER[1],ANGLEUNIT["degree",0.0174532925199433]],AXIS["longitude",east,ORDER[2],ANGLEUNIT["degree",0.0174532925199433]],AXIS["ellipsoidal height",up,ORDER[3],LENGTHUNIT["metre",1]],ID["EPSG",4979]]],ABRIDGEDTRANSFORMATION["CGVD28 height - HT2_0 to WGS84 ellipsoidal height",METHOD["GravityRelatedHeight to Geographic3D"],PARAMETERFILE["Geoid (height correction) model file","HT2_0.gtx",ID["EPSG",8666]]]]]'
Thanks in advance for any assistance!
Best regards,
Steve
Steve Riddell
Software Engineering
GeoCue Group, Inc.
520 6th Street | Madison, AL 35756 USA
Phone: 256.461.8289 | Fax: 256.461.8249
LIDAR/Drone Mapping Software & Services – True View® 3D Imaging Sensors – Image/LIDAR Data Management Solutions
www.geocue.com<http://www.geocue.com/> support.geocue.com<https://support.geocue.com/>
--
http://www.spatialys.com
My software is free, but my time generally not.
--
http://www.spatialys.com
My software is free, but my time generally not.
--
http://www.spatialys.com
My software is free, but my time generally not.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/gdal-dev/attachments/20220328/fd7290bd/attachment-0001.html>
More information about the gdal-dev
mailing list