Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo


#181 Raster Import error

Extent (1)
Henri Riihimäki

I'm currently having a problem loading same extent .tif-grids to same grid-system. In Saga there appears extradecimals that I have no idea where they come from. Also the extent tends to shift a little.

My original data is 15 x 3000 x 3000 pixel lidar dem (2 m resolution, .asc) that I have mosaiced in arcmap and exported to .tif -format. In Arc everything seems to be fine, extent and all, but when I try to read this file (and associated layers calculated from dem) weird things happend. I'm sorry if this is wrong place to address such issue, but I really can't think anyplace else and I've been stuck with this issue for a good while..


  • I'm not sure if this is a SAGA issue or a problem in ArcMap - usually there is no scope of interpretation regarding coordinates stored in *.tif files. But in order to be able to figure out the problem, we need more information (e.g. which SAGA version and which SAGA module are you using, which settings are used etc.) and some examples (i.e. figures) on how the coordinates differ. Best would be if you could attach some (small, 3x3 cell grids are enough) test files to this ticket which allow to reproduce the problem.

  • I did some further research and it appears that ESRI grid format might be a reason in itself. I've created this thread in esri forums, relating to same problem: http://forums.arcgis.com/threads/103871-Spatial-reference-changes-itself?p=370562#post370562

    Also, setting the projection and unsetting it (in ArcGIS) might cause these weird decimals. After doing this and reading these files to Saga the extent decimals seem allmost random.

    However when no projections definitions are NOT done this is how my coordinates look:

    Real coordinates ends 000
    After mosaic in ArcMap coordinates end 000.00012207 (Coordinates are same for exported .tif too)

    And in Saga the coordinates appear 001.000122y (Saga reads the center of the cell, thus 001.. and not 000.. ?)

    I'm using Saga v. 2.1.0 with default settings. (load > all files > myRaster.tif )
    (I'll suppose that is the same as using GDAL Raster import module with defaults)

  • Ok, so I think everything is fine with SAGA - any objections to close this "bug" report?

    BTW, SAGA references the grid cell center and not the the lower left grid cell area, so this might explain the reported lower left coordinate of the grid system. In case you like to investigate this in more detail, you could export your grid in ArcMap as ESRI ASCII raster - then you have all coordinates in plain text when opening the file in an editor.



Cancel   Add attachments