[Xastir] vertical lines in hi-res geotiff orthophotography

[Tom Russo]

On Tue, Jan 18, 2011 at 03:27:26AM -0500, we recorded a bogon-computron collision of the <kevin at kevinratcliff.com> flavor, containing:
> Hi all,
> 
> I'm trying to use some high-resolution (.15 meter) aerial photos in
> Xastir. The files in question work great, except at zoom levels 1 and
> 2. At those zoom levels I see lots of gray vertical lines that obscure
> the image to the point where it's no longer useful. I don't see this
> issue using DRG topo files.
> 
> Here is a link to the source files for free downloads and metadata:
> 
> http://www.indiana.edu/~gisdata/local/atterbury_muscatatuck2010.html
> 
[...]
> 
[...]
> I tested on multiple systems (two Ubuntu 10.04 and one Ubuntu 10.10,
> all installed using the Ubuntu wiki instructions for CVS and the
> UbuntuGIS PPA). One of the systems is at CVS from January 17, 2011,
> the others are a few weeks older. Two are desktops and one is a
> laptop, so the video hardware differs quite a bit; none are VMs. I'm
> about to try a fresh Fedora install to see if it does the same thing.

I would expect that this is a bug in the rendering method in Xastir, not
a system-dependent thing (so I doubt your fresh Fedora experiment will show
any improvement).

Xastir draws rectangles for each pixel using XFillRectangle, and my guess is
that the place where it calculates the size of the rectangles is getting the
calculation just a tiny bit off.  This is just a hunch, but most of the imagery
that has been used in Xastir so far is far lower resolution than these, and
it's entirely possible there's a numerical issue in the computation.  At least,
that's what this sort of thing smells like to me just from your description.

I have occasionally used 1-meter-per-pixel and 1 foot-per-pixel orthoimagery
in Xastir and not seen what you're seeing, but my guess is that we're starting
to get to edge cases for the applicability of the approximations being made
in the code.  Tidying that up could be a challenge, but given how much 
high-quality free orthoimagery is starting to be available, it'd be worth it.

> What information do you need from me to help track this down?

I'd like to look into it, but it would help tremendously if you would tell me
exactly which imagery file you're looking at, say the one with the
SAR Training Center in it.  Looks like the web site you provided has lots of
imagery files, and I'd rather download precisely the one that you're looking at.

> What debug level should I set in Xastir? Currently I'm not seeing any
> errors or warnings in the terminal.

There wouldn't be any errors or warnings if my hunch is correct, just 
miscalculation of pixel widths.  But you can see some diagnostics about what 
the tiff rendering code is doing by setting debug level 16; it might be 
enlightening, or might just be a lot of useless numbers.  There are a 
number of debug lines commented out in map_tif.c that might also be of 
interest.  But maybe not of interest to you.

-- 
Tom Russo    KM5VY   SAR502   DM64ux          http://www.swcp.com/~russo/
Tijeras, NM  QRPL#1592 K2#398  SOC#236        http://kevan.org/brain.cgi?DDTNM
 "The truth will set you free, but first it will piss you off."