Thanks for looking into this. I've been doing some more work, and during a lot of my attempts to reproject data I get "Bursa Wolf Parameter" errors. From reading along, I got the impression this is due to a lack of data about projections and this might be solved in a future release by getting better EPSG data (??).
Is there any temporary workaround to manually provide the parameters that it's looking for? > I did some experimenting yesterday and found that raster reprojection > was taking place with world+image files, but it wasn't very good, I'm > not sure why. I reprojected a map of victoria in BC_Albers to WGS_84 > and there was definate reprojection taking place, but it did not line > up well with the reprojected shapefile. I'll keep digging. > > With regards to WMS, Server side reprojection is attempted first and > if that fails raster reprojection is performed within uDig. So it > depends on the server and what reprojections take place. > > Jesse > > On 15-Mar-06, at 6:59 AM, Christopher Hammack wrote: > > > Thanks for looking into this. I'm not seeing very good results-- my > > images are simply png+worldfiles that I create georesources from. I > > have tried with and without setting the layer's CRS. > > > > I just tried using a WMS server (JPL's server listed on the test > > server > > list), and I saw good results in reprojection to Lambert Conic and > > Mercator. > > > > So am I correct in assuming that reprojection works for WMS but not > > for > > worldfile rasters? Or is the WMS server doing the projection on the > > server side (sorry, not familiar with how WMS works)? > > > >> I've spent a few hours today looking through the code and I can > >> verify that raster reprojection is happening during rendering... It > >> isn't perfect I admit but more cannot be done at the moment, at > > least > >> not until the new coverage support is finished being added to > >> geotools. I noticed now (the second read through) that you > >> mentioned calling layer.setCRS(). This isn't intended to do any > >> reprojection. Instead it is a way to declare what projection the > >> layer is in. This was provided for the cases where there was no > >> projection file or the projection file is wrong. Instead only the > >> viewportModel's setCRS will affect the rendering. > >> > >> Jesse > >> > >> On 14-Mar-06, at 11:32 AM, Christopher Hammack wrote: > >> > >>> I've read some messages recently regarding raster reprojection, but > >>> I've yet to hear confirmation as to exactly what should be working. > >>> > >>> If I create a map that is Lambert Conformal Conic (by calling the > >>> ViewportModelInternal's setCRS method), vector data (e.g. > > shapefiles, > >>> gml, etc.) look great. > >>> > >>> However, if I take a raster image, with a standard projection, > >>> create a > >>> layer from it and call setCRS on the layer, I see no reprojection. > >>> > >>> Looking at geotools code, it appears there is the capability to > >>> reproject rasters by breaking them into tiles, etc. Perhaps I have > >>> misunderstood this, however. > >>> > >>> As one might expect, the non-projected raster is roughly in the > > right > >>> area, but is considerably off. > >>> > >>> Any suggestions? Thanks. > >>> -- > >>> _______________________________________________ > >>> User-friendly Desktop Internet GIS (uDig) > >>> http://udig.refractions.net > >>> http://lists.refractions.net/mailman/listinfo/udig-devel > >> > >> _______________________________________________ > >> User-friendly Desktop Internet GIS (uDig) > >> http://udig.refractions.net > >> http://lists.refractions.net/mailman/listinfo/udig-devel > >> > >> > > > > -- > > _______________________________________________ > > User-friendly Desktop Internet GIS (uDig) > > http://udig.refractions.net > > http://lists.refractions.net/mailman/listinfo/udig-devel > > _______________________________________________ > User-friendly Desktop Internet GIS (uDig) > http://udig.refractions.net > http://lists.refractions.net/mailman/listinfo/udig-devel > > -- _______________________________________________ User-friendly Desktop Internet GIS (uDig) http://udig.refractions.net http://lists.refractions.net/mailman/listinfo/udig-devel
