[Gdal-dev] tiled images into a jp2

[Fu Chen]

Frank,
I tried to use vrt to wrap all those tiled tiffs into a virtual dataset.
and use gdal_translate to translate the vrt file into a jp2.
But maybe this dataset is too large(105G, dimension 200K*200K), the
progress is very slow, and become much more slower. For example, the
first 5% will take an hour, but the same 5% between 35%-40% will take
one day.
The kakadu's author dont recommend to use tiled jp2, for in this way we
will find artifacts besides the border of two tile. But will tiled style
speed the compress and upcompress process of the large dataset?

-----Original Message-----
From: gdal-dev-admin at remotesensing.org
[mailto:gdal-dev-admin at remotesensing.org] On Behalf Of Frank Warmerdam
Sent: Saturday, January 31, 2004 2:45 AM
To: gdal-dev at remotesensing.org
Subject: Re: [Gdal-dev] tiled images into a jp2


Fu Chen wrote:
> Hi!
> I have a very large geospatial image(105GB), which has already divided

> into equal size tiles that each one is a raw image file around 500MB. 
> I wanna translate it into a jp2 dataset by kakadu. Is it possible that

> we first create a jp2 with a tile size and then translate each one 
> into a tile struction in the jp2?

Fu Chen,

As things are implemented currently in JP2KAKDataset::CreateCopy(), what
you want isn't supported.  However, it is possible to accomplish what
you want with Kakadu.  If you are really wanting to do this, you might
want to build a custom writer on top of Kakadu just using GDAL to read
the data.

Best regards,
-- 
---------------------------------------+--------------------------------
---------------------------------------+------
I set the clouds in motion - turn up   | Frank Warmerdam,
warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent


_______________________________________________
Gdal-dev mailing list
Gdal-dev at remotesensing.org
http://remotesensing.org/mailman/listinfo/gdal-dev