What's a good free viewer for tiled ortho tiffs?
Spud
continuous image when tiled together. I am looking for a cheap (free)
viewer that can display the tiles as a seamless continuous image so I can
print, scroll and zoom over tile boundaries. For manageability I realise
that I may have to cut the existing 40MB tiles in to 4 x 10MB ones - that's
not a problem. I can also georegister these with ease.
I assume most (all?) of these tiling arrangements require a co-ordinating
file that informs the viewer which files it needs to load/display. I don't
have access to the "parent" application to most of the viewers I have seen,
so the co-ordinating file would ideally be a logically arranged text file.
I am presently using ACDSee to view the images which is OK for performance
but of course cannot dynamically tile them together. I don't need the
viewer for any other vector or raster display so it can be limited in its
functionality.
Any suggestions are welcome.
Spud
(A real couch potato)
Jonathan Buzzard
Try MapSheets Express from Erdas. Will do everything you want but print
properly. You can embed the image in a Word document or otherwise but they
seem to resample the image down by at least two as the output is awful.
JAB.
--
Jonathan A. Buzzard Email: jona...@buzzard.org.uk
Northumberland, United Kingdom. Tel: +44(0)1661-832195
B@rt
georeference.
Get version 2.0, the new 3.0 java version would not run properly on my
system...
http://www.esri.com/software/arcexplorer/index.html
Cheers,
Bart Adriaanse
The Netherlands
"Spud" <pal89de...@bigfoot.com> wrote in message
news:99145437...@dns2.netaccess.net.nz...
Jonathan Buzzard
"B@rt" <B...@rt.xl> writes:
> ArcExplorer from ESRI will do fine, if you have the .TFW world files for
> georeference.
>
> Get version 2.0, the new 3.0 java version would not run properly on my
> system...
>
> http://www.esri.com/software/arcexplorer/index.html
>
> Cheers,
>
I always found that sucked a little bit performance wise. I would go for
either ERDAS MapSheets Express or there Viewfinder program. These
calculate pyramid layers for TIFF files so perfomance is still
respectible for some time.
However in short there is no free viewer that will display TIFF's
with decent performance (i.e. use pyramid layers) and do decent simple
map compositions. In particular with a few simple shapefile layers.
Spud
few others) and appreciate the pros and cons of each now. As you say
Jonathon, the Erdas product does what it can to boost performance but on a
run of the mill machine that could still be a bit slow. For reasonable
performance it looks like a few bucks might be involved.
I might try and find a friendly local with one of the full GIS/image
products to set up an image pyramid in Mapinfo (or similar) with zoom
control of layers. I think the Mapinfo freebie reads the workspace files
and hopefully that is where the zoom control info is stored and the free
viewer can use it.
Thanks for your help again.
Cheers Spud
Jonathan Buzzard <jona...@happy.buzzard.org.uk> wrote in message
news:a4pbf9...@192.168.42.254...
Simon Dawn
<pal89de...@bigfoot.com> wrote:
>I have a collection of orthorectified tiffs (40MB each) that form a
>continuous image when tiled together. I am looking for a cheap (free)
>viewer that can display the tiles as a seamless continuous image so I can
>print, scroll and zoom over tile boundaries. For manageability I realise
>that I may have to cut the existing 40MB tiles in to 4 x 10MB ones - that's
>not a problem. I can also georegister these with ease.
You could also try the ECW format (www.ermapper.com), which stores
images into a compressed image format with very fast access. Where as
ERDAS's products struggle displaying images over a GB in size, the ECW
format is really quick, even for very large images. They have
examples of 100GB mosaics as a single image online you can see at
www.earthetc.com.
There is a free ECW compressor for images up to 500MB in size, a free
viewer you can include with images, and free plugins for most GIS
products (MapInfo has native ECW support).
If you want to create a full single-image mosaic bigger than 500MB,
you might have to look at the full ER Mapper product, which can
mosiak, color balance and compress a bunch of input images.
www.EarthEtc.com has examples.
Hope this helps
Simon
Jonathan Buzzard
Simon Dawn <nos...@nowhere.com> writes:
> On Sat, 2 Jun 2001 15:53:31 +1200, "Spud"
> <pal89de...@bigfoot.com> wrote:
>
>>I have a collection of orthorectified tiffs (40MB each) that form a
>>continuous image when tiled together. I am looking for a cheap (free)
>>viewer that can display the tiles as a seamless continuous image so I can
>>print, scroll and zoom over tile boundaries. For manageability I realise
>>that I may have to cut the existing 40MB tiles in to 4 x 10MB ones - that's
>>not a problem. I can also georegister these with ease.
>
> You could also try the ECW format (www.ermapper.com), which stores
> images into a compressed image format with very fast access. Where as
> ERDAS's products struggle displaying images over a GB in size, the ECW
> format is really quick, even for very large images. They have
> examples of 100GB mosaics as a single image online you can see at
> www.earthetc.com.
I never noticed that? Personally I found that Imagine was faster than
Ermapper for displaying images simply because of better use of pyramid
layers. Make sure they are calculated and working properly and there
is nothing faster. I am not really sure of the advantage of a single
100GB image though. Distribution via Ultrium tape only I suppose, hum.
> There is a free ECW compressor for images up to 500MB in size, a free
> viewer you can include with images, and free plugins for most GIS
> products (MapInfo has native ECW support).
The free viewer cannot however bring in two georeferenced images and
display them side by side. At least the last version I checked could not.
> If you want to create a full single-image mosaic bigger than 500MB,
> you might have to look at the full ER Mapper product, which can
> mosiak, color balance and compress a bunch of input images.
> www.EarthEtc.com has examples.
>
I am not sure that is fantastically usefull. I generally think that a
number of smaller images that are georefferenced view a viewer that
knows the right thing to do and can save image sets is better. In
particular images bigger than 650MB are more trouble than they are
worth at least in the final stage of distributing data to clients.
Bernhard Reiter
> continuous image when tiled together. I am looking for a cheap (free)
> viewer that can display the tiles as a seamless continuous image so I can
> print, scroll and zoom over tile boundaries.
Under some conditions
Orthophoto Server
http://tull.mit.edu/orthoserver/
http://mapus.jpl.nasa.gov/
might be a possible solution to your problems.
I have not checked the details, though.
Other Free Software (as in freedom) can be found at www.freegis.org.
The OSSIM project certainly aims to create a very useful
remote-sensing application, but I am not sure how ready the
application is for the task already.
Bernhard
--
Professional Service around Free Software (intevation.net)
The FreeGIS Project (freegis.org)
Association for a Free Informational Infrastructure (ffii.org)
FSF Europe (fsfeurope.org)
John D. Evans
In article <9ficcn$4e80l$1...@ID-89274.news.dfncis.de>,
Bernhard Reiter <bern...@intevation.de> wrote:
>In article <99145437...@dns2.netaccess.net.nz>,
> "Spud" <pal89de...@bigfoot.com> writes:
>> I have a collection of orthorectified tiffs (40MB each) that form a
>> continuous image when tiled together. I am looking for a cheap (free)
>> viewer that can display the tiles as a seamless continuous image so I can
>> print, scroll and zoom over tile boundaries.
>
>Under some conditions
> Orthophoto Server
> http://tull.mit.edu/orthoserver/
> http://mapus.jpl.nasa.gov/
>might be a possible solution to your problems.
I'd say the MIT OrthoServer is quite close to what you want. It
assembles seamless views in real time from individual TIFF images and
a simple ASCII index file; and outputs them in TIFF, geoTIFF, JPEG, or
PNG format.
The TIFF image extractor/mosaicker runs behind a Perl script that
accepts queries via the Web according to the OGC Web Map Service
protocol (http://www.opengis.org/techno/specs/00-028.pdf). A second
optional Perl script provides a rudimentary GUI.
The software components are available under open source terms at
http://tull.mit.edu/orthoserver ; and you can see the server in action
at the following URLs:
http://ortho.edc.uri.edu (Rhode Island, USA)
http://ortho.mit.edu/nsdi/seamless6.cgi (Boston area)
http://ortho.gis.iastate.edu (state of Iowa)
http://ortho.ftw.nrcs.usda.gov (US coast-to-coast DRGs)
[Bernhard mentioned http://mapus.jpl.nasa.gov as well -- in fact this
has a different pixel extractor running behind the Perl scripts; it
works on a single very large TIFF (218,000x95,000x6 8-bit pixels)
encoded with an experimental compression scheme. Very interesting; but
a somewhat different beast!]
Please contact me if you'd like further details on this software.
John D. Evans, Ph.D.
john....@gsfc.nasa.gov
GST, Inc. / NASA Digital Earth Office
Clay
<pal89de...@bigfoot.com> wrote:
For viewing go here:
http://www.pcigeomatics.com/freeware/freeware.html
You can load and view 80+ raster/vector formats, regardless of depth,
format, projection, resolution etc. This version does NOT print but
try again later this summer for this functionality.
Simply put, this is the best freeware you can buy.
Simon Dawn
(Jonathan Buzzard) wrote:
>> You could also try the ECW format (www.ermapper.com), which stores
>> images into a compressed image format with very fast access. Where as
>> ERDAS's products struggle displaying images over a GB in size, the ECW
>> format is really quick, even for very large images. They have
>> examples of 100GB mosaics as a single image online you can see at
>> www.earthetc.com.
>
>I never noticed that? Personally I found that Imagine was faster than
>Ermapper for displaying images simply because of better use of pyramid
>layers. Make sure they are calculated and working properly and there
>is nothing faster. I am not really sure of the advantage of a single
>100GB image though. Distribution via Ultrium tape only I suppose, hum.
Actually, the whole point about the Image Web Server
stuff is it serves big images over the net. I think it uses a
streaming wavelet technology built on their ECW stuff. So you don't
need a tape.
My impression is that Imagine is pretty slow with image display even
after building those darn pyramids for each image. And the pyramids
end up with images being bigger than the original file, ECW compresses
the images, so they are also smaller. Wavelets let you get any level
view quickly. Certainly Imaging is slower than ERmapper unless
pyramids are built, and much slower than ERmapper with ECW images.
And where as Imagine writes a disk file for just about every process,
ERMapper does it on the fly. Try viewing a PCA in real precision
using Imagine - not only is it deathly slow, but it uses about 10GB if
disk space just for a single Landsat image - ERMapper does it in real
time without temp disk!
ERMapper is much much quicker than Imaging for handling large images -
you only have to see all the ERDAS-L problems about needing GB of temp
space to see how clumsy it is.
As for speed - go to www.earthetc.com and check out those big images -
you can view 100GB images using a browser over the network faster than
Imaging can access a 1GB image from disk. And you can also view the
image over the net from WORD and Arcview.
>> If you want to create a full single-image mosaic bigger than 500MB,
>> you might have to look at the full ER Mapper product, which can
>> mosiak, color balance and compress a bunch of input images.
>> www.EarthEtc.com has examples.
>>
>
>I am not sure that is fantastically usefull. I generally think that a
>number of smaller images that are georefferenced view a viewer that
>knows the right thing to do and can save image sets is better. In
>particular images bigger than 650MB are more trouble than they are
>worth at least in the final stage of distributing data to clients.
>
>JAB.
It depends on what you are trying to do. A CD can store 600MB of
image. That is about 15GB of uncompressed imagery. So I find it
easier to put all the images together, and color balance them, then
hand them out as a single image. So it is easy to give out 15GB of
images all on a single CD, and better than trying to open lots
of images and sorting out the checkerboard problem.
Simon
Jonathan Buzzard
Well when I have done it, it only wants 7GB of temp disk space. However
what you are trying to claim is that ERmapper will bring in the file from
disk, compute a PCA and then a zoomed out view and display it faster than
Imagine can bring the zoomed out image in from disk and display it.
I really don't think this is the case. It certainly is not in my experience.
Yes to initally view the image on Imagine may be slower but once the PCA is
computed it is faster. ERmapper does appear to compute pyramid layers, it
just junks them every time you close the image.
> ERMapper is much much quicker than Imaging for handling large images -
> you only have to see all the ERDAS-L problems about needing GB of temp
> space to see how clumsy it is.
Yes and no. ERmapper and Imagine have taken different stand points. With
ERmapper they have assumed the disk space is limited and expensive. To over
come this they do everything on the fly. Imagine takes a different stand
point, disk space is getting ever bigger and cheaper so we will burn disk
space for later faster image viewing.
Now in the long run I would say that the approach taken by Imagine is always
going to win out. With IBM's "pixie dust" technology going to deliver
200GB laptop drives and 400GB desktop drives in the next two years then
the view that the disk space is cheap and nearly limitless would appear
to be the correct one.
> As for speed - go to www.earthetc.com and check out those big images -
> you can view 100GB images using a browser over the network faster than
> Imaging can access a 1GB image from disk. And you can also view the
> image over the net from WORD and Arcview.
You miss the point, a single 100GB image is entirely pointless. You can
achieve the same result with a set of edge matched georefferenced images.
>>> If you want to create a full single-image mosaic bigger than 500MB,
>>> you might have to look at the full ER Mapper product, which can
>>> mosiak, color balance and compress a bunch of input images.
>>> www.EarthEtc.com has examples.
>>>
>>
>>I am not sure that is fantastically usefull. I generally think that a
>>number of smaller images that are georefferenced view a viewer that
>>knows the right thing to do and can save image sets is better. In
>>particular images bigger than 650MB are more trouble than they are
>>worth at least in the final stage of distributing data to clients.
>>
>>JAB.
>
> It depends on what you are trying to do. A CD can store 600MB of
> image. That is about 15GB of uncompressed imagery. So I find it
> easier to put all the images together, and color balance them, then
> hand them out as a single image. So it is easy to give out 15GB of
> images all on a single CD, and better than trying to open lots
> of images and sorting out the checkerboard problem.
>
For many uses lossy compressed imagery is simply not acceptable. In
particular with any form of automated image processing such as feature
recognition, land classification if you use a compressed image you
can *never* be sure if you are detecting compression artifacts or the
real thing. Note the lots of images should be easy (old versions of
imagine where not) and usually you can save an image set which you
subsequently load in. The viewer should sort out putting all the files
in the right place based on the georeference information embedded in
the file. If it does not then it's a crap viewer.
Even still if the compressed image does not fit on a CD then it is
rather pointless. Though with the new DVD-RW drives comming on the market
at a the $1000 mark, you could reasonably move this upto the 4.7GB mark.
LBSZone
You can get the free viewer via a 12 MB download from their website
Richard Host
> Well when I have done it, it only wants 7GB of temp disk space. However
> what you are trying to claim is that ERmapper will bring in the file from
> disk, compute a PCA and then a zoomed out view and display it faster than
> Imagine can bring the zoomed out image in from disk and display it.
I just did a test here with ER Mapper. From the input Landsat image,
it took only 4 seconds from start to end to load and display
a 3 band PCA for a full Landsat scene.
The problem with Imagine is the time needed to create the temp output file.
For your 7GB output PCA analysis on a 250MB Landsat image, Imagine takes
5 or 10 minutes processing before you can view anything. But
ER Mapper processes and displays on the fly - no waiting to generate
huge temp files like Imagine does.
Imagine is like the old Unix troff document formatting softare -
it is disk input to disk output before results come out. ER Mapper
is like a desk top processing system - you get output in real time.
Once you get a handle on ER Mapper's way of working, it is horrible
going back to older products like Imagine.
> I really don't think this is the case. It certainly is not in my experience.
> Yes to initally view the image on Imagine may be slower but once the PCA is
> computed it is faster. ERmapper does appear to compute pyramid layers, it
> just junks them every time you close the image.
ER Mapper is minutes faster than Imagine when you compare total time to
process and display from input image to output display.
But if you want to compare raw display speed, by comparing
display time for Imagine pyramids against ER Mapper's
ECW, ER Mapper still wins hands down.
ER Mapper is still much faster than Imagine because
you get the results there and then without having to write a bunch
of output files, and then build pyramids on the output files.
> Yes and no. ERmapper and Imagine have taken different stand points. With
> ERmapper they have assumed the disk space is limited and expensive. To over
> come this they do everything on the fly. Imagine takes a different stand
> point, disk space is getting ever bigger and cheaper so we will burn disk
> space for later faster image viewing.
The disk savings in ER Mapper is only an side benefit. The main thing
that it is *quick* and interactive. Kind of WYSIWYG image processing.
Imagine is an older product, and shows its limitations.
And even the disk savings are considerable, especially as image sizes
are growing faster than disk space.
Back to the PCA example, with 7GB output from a single 250MB input
Landsat image. That means you need 140GB of temp space just to
process a PCA for a 20 scene Landsat mosaic.... and if you want
to say merge that with SPOT to sharpen it, Imagine resamples
the temp output file - so that means 800GB of temp disk space
(or have they fixed that dumb problem now).
I just finished up a 180GB city-wide orthophoto mosaic
for a client, using ER Mapper. I did the entire thing without any
temp disk space. When I tried a smaller mosaic a couple of
years ago with Imagine, it needed about 16 times more disk space
than the input images, for all those darn temp files - so Imagine
needs about 3TB to process this city scene (and that
is why I use ER Mapper).
> Now in the long run I would say that the approach taken by Imagine is always
> going to win out. With IBM's "pixie dust" technology going to deliver
> 200GB laptop drives and 400GB desktop drives in the next two years then
> the view that the disk space is cheap and nearly limitless would appear
> to be the correct one.
The trouble with this is that images are also getting bigger, at the
square of the pixel size. Going from 10 meter to 5 meter means
the size of an image goes from say 100MB to 400MB for the same area.
It is hard enough getting disk space for the input images, without
having to have 5x or 20x more disk space just for temp files.
ER Mapper has some rough edges and sometimes drives me mad, but
overall it is a really good tool. Much more
powerful than Imagine for heavy duty production work. One thing
I like is that each new release has new stuff as part of the
core product, and not as expensive addons.
> You miss the point, a single 100GB image is entirely pointless. You can
> achieve the same result with a set of edge matched georefferenced images.
Not really. A single 100GB image is the result of 1,000 or more orhotophotos
merged and color balanced. And it means that a GIS user can zoom from
an overview right down to streets quickly. If you output the image as
an ECW file at 50:1, the output ends up being about 2GB, which means
civil engineers can put the entire city on their notebook for field work,
as a single image.
rh
Jonathan Buzzard
rh...@hotmail.com (Richard Host) writes:
> jona...@happy.buzzard.org.uk (Jonathan Buzzard) wrote in message news:
>
>
> ER Mapper is minutes faster than Imagine when you compare total time to
> process and display from input image to output display.
>
> But if you want to compare raw display speed, by comparing
> display time for Imagine pyramids against ER Mapper's
> ECW, ER Mapper still wins hands down.
> ER Mapper is still much faster than Imagine because
> you get the results there and then without having to write a bunch
> of output files, and then build pyramids on the output files.
You are compairing apples with pairs and saying apples are better. It's
not the same thing. Using the same sort of compression Imagine is faster
at displaying images than ERmapper. Note that when you deliver an image
to a client this is *all* that is important. They don't give a monkey
how the image was generated, how much disk space it needed for generation,
or anything else.
By the way it occurs to me that as to view the whole image in ERmapper is
going to require *exactly* the same amount of data to be processed as
Imagine takes then I strongly suspect total time and total time are
very similar in both products.
>
> Back to the PCA example, with 7GB output from a single 250MB input
> Landsat image. That means you need 140GB of temp space just to
> process a PCA for a 20 scene Landsat mosaic.... and if you want
> to say merge that with SPOT to sharpen it, Imagine resamples
> the temp output file - so that means 800GB of temp disk space
> (or have they fixed that dumb problem now).
You need to examine your workflow, and pick a method more suitable for
the product at hand. You however do not need that amount of temp disk
space in my experience.
> I just finished up a 180GB city-wide orthophoto mosaic
> for a client, using ER Mapper. I did the entire thing without any
> temp disk space. When I tried a smaller mosaic a couple of
> years ago with Imagine, it needed about 16 times more disk space
> than the input images, for all those darn temp files - so Imagine
> needs about 3TB to process this city scene (and that
> is why I use ER Mapper).
Hum, not my experience.
>> Now in the long run I would say that the approach taken by Imagine is always
>> going to win out. With IBM's "pixie dust" technology going to deliver
>> 200GB laptop drives and 400GB desktop drives in the next two years then
>> the view that the disk space is cheap and nearly limitless would appear
>> to be the correct one.
>
> The trouble with this is that images are also getting bigger, at the
> square of the pixel size. Going from 10 meter to 5 meter means
> the size of an image goes from say 100MB to 400MB for the same area.
> It is hard enough getting disk space for the input images, without
> having to have 5x or 20x more disk space just for temp files.
>
> ER Mapper has some rough edges and sometimes drives me mad, but
> overall it is a really good tool. Much more
> powerful than Imagine for heavy duty production work. One thing
> I like is that each new release has new stuff as part of the
> core product, and not as expensive addons.
>
>> You miss the point, a single 100GB image is entirely pointless. You can
>> achieve the same result with a set of edge matched georefferenced images.
>
> Not really. A single 100GB image is the result of 1,000 or more orhotophotos
> merged and color balanced. And it means that a GIS user can zoom from
> an overview right down to streets quickly. If you output the image as
> an ECW file at 50:1, the output ends up being about 2GB, which means
> civil engineers can put the entire city on their notebook for field work,
> as a single image.
>
For heavens sake, you just don't seem to get this point. A single 100GB
image is pointless. You can achieve *exactly* the same results with a
series of smaller more managable images that are georeferenced with
a viewer that knows how to display the images side by side and will allow
you to save an snapshot of the images loaded into the viewer so you
don't have to load them individually every time.
Finally I note that you have in the final step squashed it all into
an output image that needs no futher processing.