[Qgis-user] WMS Image Projection

[Andreas Neumann]

Hi,

The approx 1.5m offset indicates that QGIS doesn't take a necessary 
grid-shift file into account. The grid-shift file is a separate file 
that defines the displacement vectors of coordinate systems with 
different datums.

As a European GIS user I am not comfortable with North-American CRS, but 
looking the two CRS up at epsg.io I can see that

http://epsg.io/?q=26910

vs

http://epsg.io/?q=3157

EPSG 26910 claims to have an accuracy of 4 m while EPSG 3157 claims to 
have an accuracy of 1m.

There might be a grid shift file necessary between the two CRS (I am 
personally unsure if that is the case here).

Since both, the project and the WMS are in the same CRS, I would assume 
that the offset problems occur with the parcel shapefile.

Do you know a local GIS expert in your region who knows if the 
transformation between EPSG 26910 and 3157 requires a grid-shift file? I 
would assume yes, because at epsg.io it indicates that the datums of the 
two CRS are slightly different: NAD83(CSRS) (http://epsg.io/4617) vs 
NAD83 (http://epsg.io/4269).

 From there - and if we have a proper corresponding grid-shift file - we 
can help you set up the transformation in QGIS.

Greetings,

Andreas

On 2021-04-08 10:24, Springfield Harrison wrote:

> Hi Andreas,
> 
> Thanks for the reply, I'll try to clarify:
> 
> * Version is 3.16.4
> 
> * The two municipal sources are being viewed in their own web windows 
> and are displaying the same data
> 
> * They are the source of the property lines for my GIS (downloaded 
> shapefile) and the orthophoto (WMS link, not resident TIFF or other)
> 
> * The property lines are EPSG:3157 - NAD83(CSRS) / UTM zone 10N - 
> Projected, and from the municipal source.  I've attached the shapefile 
> if you have the time.
> 
> * In QGIS, the ortho is loaded from the WMS source
> 
> * 
> https://mapservices.crd.bc.ca/arcgis/services/OrthoImagery2019/MapServer/WMSserver
> * EPSG:26910 - NAD83 / UTM zone 10N - Projected
> 
> * The project CRS is EPSG: 26910 but it doesn't seem to change anything 
> by changing that or the source projections (at least within versions of 
> NAD83 UTM Zone 10N)
> 
> * Thus, the same information displays differently online _versus_ in 
> QGIS.
> 
> * The shift seems to be 1.5 m +/-, quite a lot for what I'm trying to 
> do (building permit with tight setbacks).
> * These are commonly used projections for our neck of the woods
> * The two images I sent earlier depict the differences; in QGIS, 
> earlier orthophotos show a similar or worse displacement.  Not so on 
> the municipal web viewer.
> 
> * I've read up on projections in the tutorial and Help file, no 
> revelations there, at least that I could see.
> 
> Thanks very much for any thoughts you may have.
> 
> -----
> Cheers, Spring
> 
> On 08/Apr/2021 00:47, Andreas Neumann wrote:
> 
> Hi,
> 
> Are there two different CRS involved here? Are the vectors (parcel 
> boundaries) in a different CRS than the orthoimages? This might be a 
> problem of a missing grid-shift file in QGIS.
> 
> It is also unclear for me if you load your Orthoimages locally (from 
> tiff-files as well) or from WMS (or both) and only one of them is 
> shifted?
> 
> Can you also add information what QGIS version with what proj version 
> you use? You can find this information in the about dialogue in the 
> help menu.
> 
> Greetings,
> 
> Andreas
> 
> On 2021-04-08 02:37, Springfield Harrison wrote:
> Hello All,
> 
> Re obtaining orthophotos from a WMS server.
> 
> Locally we have orthophotos that can be brought into QGIS using the WMS 
> service.  They are also viewable on the local government web map 
> service.  In both cases I can compare them to the property lines which 
> are from the same source.
> 
> On the government web pages (2), the orthophotos and the lot lines 
> appear to coincide very well.  In QGIS, the same two layers are quite 
> misaligned.
> 
> One would suspect a CRS problem but changing the CRS of the Project and 
> the WMS image in QGIS has no effect.  Even followed the tutorial to no 
> effect.
> 
> I'm attaching screen shots of the two results.  Ant thoughts would be 
> appreciated, thank you . . . .
> 
> -----
> Cheers, Springfield Harrison
> 
> _______________________________________________
> Qgis-user mailing list
> Qgis-user at lists.osgeo.org
> List info: https://lists.osgeo.org/mailman/listinfo/qgis-user
> Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20210408/72e94023/attachment.html>