Select to view content in your preferred language

Ortho photos in a geodatabase

1833
4
06-04-2010 08:28 AM
DonFreeman
Emerging Contributor
Hello, I am just beginning to explore geodatabases. I have a 16 gig orthophoto mosaic that I would like to put into my file geodatabase. The file format is ecw. I was wondering if that is possible. My attempt to import it ran for about 30 minutes and showed no progress when I cancelled it. Should I let it run longer? After I cancelled it , it looked like the structure had been added but there was no data. Maybe the file is too big? Any advice would be appreciated.
Thanks
0 Kudos
4 Replies
RyanStovern
Regular Contributor
I would let it run longer. Since it is a very large file. I have moved 50gb images into an FGDB and it took about 5 hours to complete. I would also if possible import the image into a raster dataset or raster catalog within the FGDB.
0 Kudos
regislee
Emerging Contributor
just out of curiosity,  does anyone know what type of performance hit (if any) comes from storing such large raster dataset in the geodatabase?  I'm not sure whether or not to keep the images as raster files or if i should put them in the GDB.  Any suggestions?
0 Kudos
DonFreeman
Emerging Contributor
I have given up on this for now. I found that the computing horsepower needed just to build the geodatabase is more than reasonable available.
0 Kudos
CurtisRuck
Occasional Contributor
You can load it into an Unmanaged Raster Catalog which would leave the original image intact, and only create a pointer to it.  Otherwise the performance is theoretically faster as the Managed geodatabase format tiles out the whole Ortho into much smaller tiles (default 256x256 in ArcGIS 10, 128x128 in earlier versions).  So the processing it is doing is this tiling, which can take forever.  Depending on the original format (not sure about ECW) the performance gain is huge because the entire image isn't needed to be read.  It also recompresses the tiles from ECW format into its own internal format, which is based on GeoTIFF i believe.
0 Kudos