I used a couple of time Suse linux and now moved on to Ubuntu Hoary
Hedgehog....i like it and i decided i will continue with it. I tryed to install
GRASS with apt-get and the installer started to install version 5.0.3 . I would
like to install version 6.0...so .....is anyone running GRASS 6.0 on
Hoary.....and if....can you supply the repository or some download intructions?
I know that there are availbale Debian binaries but installing trough apt solves
the problem of dependecies..so it is comfortable to intall GRASS within apt(for
me as I am new to Ubuntu)...
Thanks
I used a couple of time Suse linux and now moved on to Ubuntu Hoary
Hedgehog....i like it and i decided i will continue with it. I tryed to install
GRASS with apt-get and the installer started to install version 5.0.3 . I would
like to install version 6.0...so .....is anyone running GRASS 6.0 on
Hoary.....and if....can you supply the repository or some download intructions?
I know that there are availbale Debian binaries but installing trough apt solves
the problem of dependecies..so it is comfortable to intall GRASS within apt(for
me as I am new to Ubuntu)...
the Hoary packages are still at 5.03. i am using successfully the DebianGIS repositories (this is the home page)
on my hoary it works quite well, though it's the 6.0.0beta2 version. I couldn't manage to get qgis working due to an outdated libgdal version in the Ubuntu repository(1.25 - qgis 0.6.2 requires 1.26 if I am right).
version 6.0.0 comes with ubuntu 5.10 breezy badger (release scheduled for october) that is already available in the repositories but definitely too much unstable imho.
anyway a good starting point would be this page
www.ubuntulinux.org/wiki/UbuntuGIS
regards, Stefano
--
Stefano Costa
Visita il mio sito: http://www.iosa.it - Software Open Source per l'Archeologia
GnuPG Key ID: 1024D/0xD0D30245
Jabber: archeosteko@jabber.linux.it
Linux Registered User #385969 counter.li.org
Im now using i.rectify to gereference some maps, but i dont understand the difference between using a 1,2 or 3 grade polynomial. I see that if i want to use the 3 grade polynomial i have to mark 10 points (or more), is a 3 grade polynomial better (in precicion and accuracy) than a 2 or 1-grade polynomial? If i use a a 1 grade polynomial is not the same if i use a 3 grade with the same number of point? where is the difference?
Im now using i.rectify to gereference some maps, but i dont understand
the difference between using a 1,2 or 3 grade polynomial. I see that
if i want to use the 3 grade polynomial i have to mark 10 points (or
more), is a 3 grade polynomial better (in precicion and accuracy)
than a 2 or 1-grade polynomial? If i use a a 1 grade polynomial is
not the same if i use a 3 grade with the same number of point? where
is the difference?
In the 1-d sense, it is the difference between fitting a regression
line, fitting a simple curve, and fitting a complex curve to the data.
Depending on your map projections and data, simple rotation and scaling
may be fine or may give very poor results. If so, you need to use a
higher order polynomial fit.
Well, the rectification works, it's the region setting which is busted
so you have to use the -c flag and figure out your region / resolution
by hand. Maybe disable it in the 6.0.1 branch (or add warning/docs)
but leave as is in 6.1-cvs. If it is commented out I fear it will be
forgotten and not fixed for years.
BTW that i.rectify is being discussed. Could somebody please tell me how the
3 orders of polynom in i.rectify compare to the following 5 rectification
methods used in SuperEditPro? Which order of polymnon is an equivalent of
which method (if it is that simple)? I need to know it in order to be able
to compare outputs of i.rectify and SuperEditPro. I also heard of other
commercial programs which use the same naming convention and would like to
know how i.rectify's "orders" compare to this "methods".
Rectification methods listed in the SuperEditPro manual:
Isotropic linear (Helmert) model - requires at least two correction vectors.
x' = a10 + a1 x + a2 y
y' = a20 - a2 x + a1 y
Anisotropic linear (affine) model - requires at least three correction
vectors.
x' = a10 + a11 x + a12 y
y' = a20 + a21 x + a22 y
Bilinear model - requires at least four correction vectors.
x' = a10 + a11 x + a12 y + a13 xy
y' = a20 + a21 x + a22 y + a23 xy
Biquadratic model - requires at least nine correction vectors.
x' = a10 + a11 x + a12 y + a13 xy + a14 x2 + a15 y2 +
a16 x2y + a17 xy2 + a18 x2y2
y' = a20 + a21 x + a22 y + a23 xy + a24 x2 + a25 y2 +
a26 x2y + a27 xy2 + a28 x2y2