GPS - Calibration of maps...

Author

Calibration of maps...

Marcus Fox

2007-08-27, 12:33 pm

I've a query about the calibration of maps - in OziExplorer or Memory map
where the lines of latitude and longtitude do not follow straight lines but
a curve, for example small scale maps, such as one covering the whole world
or a continent.

Would the software get confused if I only used the minimum number of
calibration points or display my location innacurately? Or would the
calibration take into account that the lines of latitude and longtitude
curveeee depending on the scale?

Marcus

Jack Erbes

2007-08-27, 10:33 pm

Marcus Fox wrote:
> I've a query about the calibration of maps - in OziExplorer or Memory map
> where the lines of latitude and longtitude do not follow straight lines but
> a curve, for example small scale maps, such as one covering the whole world
> or a continent.
>
> Would the software get confused if I only used the minimum number of
> calibration points or display my location innacurately? Or would the
> calibration take into account that the lines of latitude and longtitude
> curveeee depending on the scale?
>

The calibration process normally has you identify the projection used
for the map, that lets the software deal with the convergence issues.
If you can come up with five calibrations points, like one one each
quadrant and one near the center, that would be the ideal situation. Or
so I think anyway, I not done much calibration of map images.

If I remember right, there is a good tutorial on calibrating maps in the
Ozi Explorer trial version.

Jack

John Tserkezis

2007-08-27, 10:33 pm

Newsgroups: alt.satellite.gps
X-Trace: 1188267664 14405 211.31.52.191
User-Agent: Thunderbird 2.0.0.6 (Windows/20070728)
MIME-Version: 1.0
In-Reply-To: < qMKdnWFkp70ZmE7bnZ2d
nUVZ8sKlnZ2d@bt.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Original-Bytes: 2951
Xref: number1.nntp.dca.giganews.com alt.satellite.gps:99696

Marcus Fox wrote:

> I've a query about the calibration of maps - in OziExplorer or Memory map
> where the lines of latitude and longtitude do not follow straight lines but
> a curve, for example small scale maps, such as one covering the whole world
> or a continent.

> Would the software get confused if I only used the minimum number of
> calibration points or display my location innacurately? Or would the
> calibration take into account that the lines of latitude and longtitude
> curve depending on the scale?

I can only speak for Oziexplorer, it's the only one I've worked with, and so
other cartography software may be different:

Providing the map image is correctly rectified, you only need two
calibration reference points.

If the image is not rectified (slanted), then you need three calibration points.

If the image is distorted in some way, then you need more points, and Ozi
does it's best to "force fit" and twist the calibration to suit.
It's in this case most people get themselves into trouble - they *think*
they're doing the right thing, and succeed in only stuffing things up further.

I've seen people calibrate maps with 10+ points, where only two were needed
for marginally better calibration accuracy.
Armed with some moderate cartography knowledge, two points are much faster
to enter, less likely to get things wrong than with 10+ points, and slightly
better results to boot.

If you're going to be doing any reasonable amount of calibrations, it's best
if you get a handle on cartography basics, at the very least in the projection
you'll be dealing with.
While I can't say doing this is hard, it's not "easy" either, as in nothing
is handed to you on a silver platter- and if you're expecting simple
step-by-step instructions to work for everything, you're going to be disappointed.

Best I can suggest, is to read up on as much as you can on the projection
your maps are in, and then experiment, and see how things are twisted when you
move numbers around. Some projections are dead easy, others have more options
that bend things quite substantially.
--
Linux Registered User # 302622
<http://counter.li.org>

Gordon Hamm

2007-09-11, 10:33 pm

On 2007-08-27 10:09:56 -0700, "Marcus Fox"
<please-reply-via-newsgroup-th@-i-posted-to.com> said:

> I've a query about the calibration of maps - in OziExplorer or Memory map
> where the lines of latitude and longtitude do not follow straight lines but
> a curve, for example small scale maps, such as one covering the whole world
> or a continent.
>
> Would the software get confused if I only used the minimum number of
> calibration points or display my location innacurately? Or would the
> calibration take into account that the lines of latitude and longtitude
> curveeee depending on the scale?
>
> Marcus

When the lines are curved you have to know the projection of the map.
The projection could be Lambert Conformal Conic, Polyconic American
etc. Then there is settings for these projections which could be
Central Meridian,
Latitude 1, Latitude 2, Origin Latitude. These are all settings in
OziExplorer's calibration menu.
If the maps are scans of paper maps, the projection could be on the original.
I normally use 4 calibration points for any map. Top left, Top right,
Bottom left and Bottom right. I have calibrated
many maps from scans over the years.
To know if your calibration is accurate open the map and lay a grid
over the map. The grid should match the grid on the map. On the taskbar
go Map/Grid Line Setup select either Lat/Long or Other Grid which is
Metric. Select the Line Interval and put a check grid on. If you have
selected the proper Line Interval the grid will overlay the grid on the
map, there can be some variance but not much if you have calibrated the
map properly.

LinkBot