[Qgis-developer] On-The-Fly CRS Transformation Does Not Apply +no_off Parameter

[Hilmy Hashim]

Just did a test using GDAL 1.10.0 on Windows XP. It looks like ogr2ogr is
now giving the correct results with both +no_uoff and +no_off.

Now we'll wait for QGIS to pick up the new GDAL.


On Mon, Apr 29, 2013 at 12:05 AM, Andre Joost <andre+joost at nurfuerspam.de>wrote:

> Am 28.04.2013 07:21, schrieb Hilmy Hashim:
>
>> Hello all,
>>
>>
>> This query is about a discrepancy between projection results using the
>> *Custom
>> CRS Test* and *on-the-fly* transformation, and not about RSO per se.
>>
>>
>>
>> The Proj4 parameters for GDM2000 / Peninsula RSO (EPSG:3375) as given by
>> epsg_tr.py (GDAL 1.9.2) are:
>>
>> +proj=omerc +lat_0=4 +lonc=102.25 +alpha=323.0257964666666 +k=0.99984
>> +x_0=804671 +y_0=0 +gamma=323.1301023611111 +ellps=GRS80 +units=m +no_defs
>>
>> This Malaysian projection requires the +no_off or *+no_uoff* flag (both
>>
>> works) for rotation at the natural origin as of proj 4.8. To test, I
>> created a custom CRS based on GDM2000 / Peninsula RSO adding this flag.
>> However, there seems to be some discrepancy between the on-the-fly derived
>> results and the Test in the Custom CRS definition dialog.
>>
>>
>> I set the project CRS as the custom CRS I created and added the test point
>> in WGS84.
>>
>> Test points: *1.827764875* deg N, *103.640259872222* deg E (WGS84
>> EPSG:4326)
>>
>> Expected results: Northing *202,122.062* m, Easting *627,096.132* m
>> (EPSG:3375)
>>
>> *The test points and projected coordinates are taken from the GDM2000
>>
>> Technical Manual published by the Department of Survey and Mapping
>> Malaysia.
>> *
>>
>> *Test results (standard EPSG:3375)*:
>>
>>
>> cs2cs — -240332.04, 958936.71
>>
>> QGIS Master (13d35ac) On-the-fly derived: Y = -240,332.0326, X =
>> 958,936.7057
>>
>> QGIS Master (13d35ac) Custom CRS Test:* -240,332.0326* N, *958,936.7057* E
>>
>>
>> QGIS 1.8 On-the-fly: projected coordinates not derived
>>
>> QGIS 1.8 Custom CRS Test: -240,332.0326 N, 958,936.7057 E
>>
>> *Test results (with +no_off flag)*:
>>
>> cs2cs — 202122.06,  627096.13
>>
>> QGIS Master (13d35ac) On-the-fly derived: Y =* -240,332.0326*, X = *
>> 958,936.7057*
>>
>> QGIS Master (13d35ac) Custom CRS Test: *202,122.0625* N, *627,096.1317* E
>>
>>
>> QGIS 1.8 On-the-fly: projected coordinates not derived
>>
>> QGIS 1.8 Custom CRS Test: 202,122.0625 N, 627,096.1317 E
>>
>>
>> It looks like the Custom CRS Definition Test dialog is giving the correct
>> result, but the on-the-fly transformation is ignoring the +no_off flag.
>>
>>
>> Is the on-the-fly transformation using a different transformation process
>> and reading the proj4 parameters differently, or am I going about this
>> wrongly?
>>
>
> I made some tests with the similar Michigan omerc projections and ran into
> the same problem:
> http://trac.osgeo.org/gdal/**ticket/4910<http://trac.osgeo.org/gdal/ticket/4910>
>
> It looks like it is a GDAL problem.
>
> HTH,
> André Joost
>
> ______________________________**_________________
> Qgis-developer mailing list
> Qgis-developer at lists.osgeo.org
> http://lists.osgeo.org/**mailman/listinfo/qgis-**developer<http://lists.osgeo.org/mailman/listinfo/qgis-developer>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-developer/attachments/20130502/41ff3f38/attachment.html>