<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Dec 22, 2015 at 9:01 AM, Even Rouault <span dir="ltr"><<a href="mailto:even.rouault@spatialys.com" target="_blank">even.rouault@spatialys.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">Le mardi 22 décembre 2015 14:38:30, Aaron Boxer a écrit :<br>
> On Tue, Dec 22, 2015 at 4:34 AM, Jukka Rahkonen <<br>
><br>
> <a href="mailto:jukka.rahkonen@maanmittauslaitos.fi">jukka.rahkonen@maanmittauslaitos.fi</a>> wrote:<br>
> > Aaron Boxer <boxerab <at> <a href="http://gmail.com" rel="noreferrer" target="_blank">gmail.com</a>> writes:<br>
> > > This feature would be quite involved: one would have to gather all of<br>
> > > the<br>
> ><br>
> > layers for<br>
> ><br>
> > > all of the code blocks in all of the precincts overlapping that<br>
> ><br>
> > sub-tile,plus the immediately surrounding precincts, run the inverse<br>
> > entropy<br>
> > coder to get the wavelet coefficients, and then do an inverse wavelet<br>
> > transform,with this incomplete set of coefficients.<br>
> ><br>
> ><br>
> > I do not claim that I understand GIS but I have a feeling that precincts<br>
> > are<br>
> > designed for a fast ROI access. From<br>
> > <a href="http://www2.engr.arizona.edu/~bilgin/publications/SPIE2004_2.pdf" rel="noreferrer" target="_blank">http://www2.engr.arizona.edu/~bilgin/publications/SPIE2004_2.pdf</a><br>
> ><br>
> > "A precinct is a collection of code-blocks representing some finite<br>
> > spatial extent at some resolution."<br>
> ><br>
> > Probably gathering the surrounding precincts is unnecessary. Author of<br>
> > Kakadu seems to dislike tiles and it is no wonder that with Kakadu<br>
> > precincts<br>
> > perform well even with archived documents<br>
> > <a href="http://www.digitizationguidelines.gov/still-image/documents/Martin.pdf" rel="noreferrer" target="_blank">http://www.digitizationguidelines.gov/still-image/documents/Martin.pdf</a>.<br>
><br>
> Thanks, that paper was an interesting discussion of optimal encoding<br>
> settings in the archive world.<br>
><br>
> What are typical j2k encoding settings in geo-spatial use case ?<br>
<br>
</div></div>That must depend vastly on the producer.<br>
<br>
A few examples:<br>
<br>
You can download Sentinel-2 products at :<br>
<a href="https://scihub.copernicus.eu/s2/#/home" rel="noreferrer" target="_blank">https://scihub.copernicus.eu/s2/#/home</a><br>
up to 10980x10980 images, 12 bit, with 2048x2048 tiles<br>
<br>
Pleidas products :<br>
<a href="http://www.geo-airbusds.com/en/23-sample-imagery" rel="noreferrer" target="_blank">http://www.geo-airbusds.com/en/23-sample-imagery</a><br>
Bigger dimension that S-2. Also 12 bit, with 2048x2048 tiles<br>
<br>
There are likely sources that are single tiled.<br></blockquote><div><br><br></div><div>Thanks. These sites will be very useful.<br><br></div><div>Is it standard to add TLM and PLM markers? Because these make fast random access<br></div><div>possible.<br><br><br></div><div> <br></div></div><br></div></div>