[EMAIL PROTECTED] wrote:
> ...I discovered that if I entered more than 3
> control points, the accuracy of the digitizer-to-mapinfo mapping was seriously
> degraded (both in the numeric calculations and in comparing puck location with
> the map on the screen). The more control points that were included, the
> greater the location errors. The best accuracy always occurred with only 3
> control points.
>
> why?
Hi All!
As Ole Gregor pointed out, 3 control points is no control, only the
minimum the algorithm needs to do its job. With 4 points you get one
extra, I'd guess you can call it oversampling.
With each added control point, the accuracy should increase.
Why does it not? Well, I'd like to put forward a theory.
Perhaps mr Figuers is trying to digitize from one projection into
another?
Let's consider a flat drawing (non-earth) as opposed to the curved earth
(projected coordinates).
A flat triangle can be placed with all vertices touching the sphere, and
perhaps also a 4 cornered shape. But as you add more control points, the
task gets more difficult, hence a greater overall error.
I frankly admit that I'm no geodetic expert, and perhaps I'm all wrong
here, and would gladly welcome other explanations.
G'night from Mats.E
--
:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
:Email/ [EMAIL PROTECTED] ICQ#9517386 :
:---------------------------------------------------------------------:
:Mail/ GISKRAFT, Mats Elfstroem, Vaepplingv 21, SE-227 38 LUND, SWEDEN:
:Phones: +46 46 145959, +46 70 595 3935 :
:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
----------------------------------------------------------------------
To unsubscribe from this list, send e-mail to [EMAIL PROTECTED] and put
"unsubscribe MAPINFO-L" in the message body, or contact [EMAIL PROTECTED]
