Kompass map / UTM calibration

FEDE · 3. August 2003

Trying to calibrate scanned Kompass map K617. Maps shows UTM grid (red) + geographical coordinates (blue). Datum is said to be International (red) and European 1950 (blue).

After calibrating the map I notice that TTQW's grid in UTM always overlaps with the red map's UTM grid, but has a shift from the geo blue coordinated.

What is the reason of the shift?

What are the best settings to use for calibrating Kompass maps?

jockel · 3. August 2003

Hi,

one possibility:
probably you have calibrated in ED50, lat/lon. You switch to UTM grid,
but you forget, that you have to set Datum to "International" to have the
TTQV UTM grid matching the red map UTM grid.

I think the error must be something like this if you see a constant error over the entire grid !

Jo

FEDE · 3. August 2003

I tried calibration with International, WGS84 and Eu1950 and using UTM (the UTM tab).

In fact collimation of TTQW UTM grid with map's grid does not change much with any of the 3 DATUM above, it stays close.

Now, if change TTQV grid unit to DD:MM:SS, I see that the TTQW grid does not match the DD:MM:SS. Let me underline the scanned map HAS TWO coordinate systems, UTM and DD:MM:SS and I compare DD:MM:SS with DD:MM:SS in this case.

I tried most of the combination of the 3 datums. DD:MM:SS shift changes, but not very much either depending on the combination. Ocasionally the shift may be close to 0, but it is different for each of the A4 chunk of the map.

Why do I have this shift (I pretend to say that the accuracy of setting the reference point at calibration time was similar).

jockel · 3. August 2003

Zitat von "FEDE"

I tried calibration with International, WGS84 and Eu1950 and using UTM (the UTM tab).
In fact collimation of TTQW UTM grid with map's grid does not change much with any of the 3 DATUM above, it stays close.

FEDE,
just to make certain that we understand each other, I did not talk about
calibrating the map in TTQV using different datums but the same grid, but
about changing the datum in "units" after you select the respective grid.

You said, UTM is supposed to be "international"(I think this sould read
WGS84!!), lat/lon is ED50. So, for example, you have to calibrate
in "International" if you use the UTM grid and when you switch in units
to "lat/lon" grid, you also have to set the datum to "ED50". You have to
follow these rules, or the relation between these grids is not correct.
Calibrating for example in UTM using whatever other datum, will always
display you a grid close to the given UTM grid, but the positioning will be wrong.

If I understand you right, you have scanned this map in portions of A4?
Now calibrate all these chunks using UTM, WGS84.
Now open the map, set units to "lat/lon", set Datum ED50. If the grids
on the map are printed correctly and the selected datum is right, the grids
should match.

If one of the datums is wrong, if the grid is not printed correctly on the
map, you will see a shift.

It is always better to compare your calibration against a recorded track
to check for correct calibration, printed grids or datum given may
be wrong.

What region is the map?
What scale ?
What age?

Maybe I can send you a track to check the calibration.

FEDE · 4. August 2003

The map is a topo map Kompass 1:25000, blatt K617. It covers UTM 33 728000, 5173000 to 32 283000, 5150000 on one side. It should be one of most recent reprints, as it has both Lat/lon and UTM. Older ones did not have UTM grid.

It has a red UTM grid, apparently referred to Datum International

It also has a DD:MM:SS grid, apparently referred to 'ED50, conform to UTM projection'. It ranges from 12:00:00, 46:40:00 to 12:10:00, 46:28:00

I said apparently beacause the legend is not too explicit about that. The 'International speheroid' and 'International spheroid ED50' datums are BOTH clearly mentioned in different colors, but it is not too clear what is associated with what, so that I assumed color coding of legend and grid should match. It could also be an artifact of the translation. WGS84 is not mentioned, but I tried it anyhow.

I gave you coordinates of top left and right bottom of the map. The 2 pair of points do not match, they are just the grid edges for each coordinate system.

I did the following:
1) scanned several A4 chunks of the map with flatbded scanner at 150dpi
2) Calibrated each chunk with UTM and datum International
3) Open map and set units to DD:MM:SS, ED50. The TTQW grid is now shifted from map's one.
4) I repeated the above with 8 chunks, then tried different datum combinations (i.e. calibrate with Int., display with ED50, Int, WGS84, then calibrate with WGS84, display with ED50, ....etc...
5) The map's grid is most always shifted, only occasionally with some combinations, they come closer. The thing that I found most puzzling is that I could not find an optimal combination, possibly the right one. Each A4, though I did not tried fully exaustively on the 8 sheets, may have a different 'optimum' combination of datums. On the other hand I noticed that somehow using same datum (WHATEVER it was) for calibration and display gives closer result on average. I guess this is an atrifact due to the small scale of the map and the way it was calibrated, but at this point I am quite confused...

The shift is relatively small, from 0 to 300 meters (measured by TTQW on the screen), but why is it there ? What is the optimal Datum for calibrating in UTM the Kompass topo serie ? And how can I solve the puzzle of matching the Lat/lon grid?

jockel · 4. August 2003

Zitat von "FEDE"

The map is a topo map Kompass 1:25000, blatt K617. It covers UTM 33 728000, 5173000 to 32 283000, 5150000 on one side.

FEDE,
one of these points is at the western border of Lake Geneve, one is
somewhere in Hungary. Better give the name of the map, so I know the
region we are talking about. I am not familiar with the map number you
have given.

"international" is no datum, but rather an ellipsoid.
ED50 is a datum, which uses - like many other datums do - the
ellipsoid "International". So it seems to me, that the only complete datum
information for a grid you have is ED50 for lat/lon. So, to my opinion,
best is to use these values for calibrating. Then check with a track about
accuracy.

Reasons why grids may not match in comparison may be:
- there is no correct datum known for the grid
- faulty calibration
- the grid is simply printed to the map wrong

It is hard to impossible to compare "non matching grids" against each
other as long as you have nothing objective to compare to (track).

If calibrating a map, there is no choice of using this or that datum with a
grid, first, the grid has to be printed accurately, second, you need to know
the right parameters, which should be given by the manufacturer,
specially with topo maps.

You mention the UTM Zone to switch within the area of this map?
In this case you should calibrate using lat/lon, as it is not possible to
calibrate a map in TTQV using different UTM zone numbers in one set of
cal-points ment for one map-chunk.

FEDE · 4. August 2003

Yes, I mistyped and inverted the zones. You should read It covers UTM 32 728000, 5173000 to 33 283000, 5150000 on one side. That covers Dolomiti Ampezzane.

I have only one zone within a single chunk.

I cannot calibrate much on lat/lon as the grid is 5 min wide and it is not practical for calibrating efficiently all the A4 chunks of the map.

I made an exercise on a single corner and calibrated same image both with lat/lon and UTM using ED50. Now the displayied TTQV with ED50 fits oblisously the map's grid used for calibration but not the other one, either UTM or lat/lon.

As I understand from you the map datum should be most likely and only ED50, I also recalibrated all the previous chunks in ED50 + UTM. Like this I have in average a 50 meter shift from the lat/lon grid.

The shift is not much, but what is the reason of it (the intrinsic error of calibrating this way? something wrong in the followed procedure? ...?) and what calibration method is supposed to be more accurate (UTM vs. lat/lon), assuming that grid on paper is correctly printed and labeled?

jockel · 4. August 2003

Zitat von "FEDE"

I made an exercise on a single corner and calibrated same image both with lat/lon and UTM using ED50. Now the displayied TTQV with ED50 fits oblisously the map's grid used for calibration but not the other one, either UTM or lat/lon.?

FEDE,
you can calibrate using the lat/lon grid of a map using whatever datum,
the lat/lon grid displayed in TTQV using this datum will always fit to the
grtid of the map.
Switching the lat/lon calibrated map to UTM and setting the datum of the
UTM grid will only match the UTM grid printed if the initially selected
Datum for lat/lon calibration and the datum switched to when switching to
UTM are correct. If they are not you will see a shift.

Zitat

As I understand from you the map datum should be most likely and only ED50?

No! Basically, trying not to go too indepth this subject: a map has no datum.
The grid printed on a map has a datum. You may have printed
on one and the same map 3 different lat lon/grids in ED50/WGS84/Rome1940,
plus 3 more UTM grids using other datums (nice mess

Zitat

The shift is not much, but what is the reason of it (the intrinsic error of calibrating this way? something wrong in the followed procedure? ...?)

The shift may be "0" despite the fact, that the calibration may be
completely wrong. There is no guessing of datums. Either you calibrate
in the known/correct datum or you may have errors in positioning. You
need a track to judge this. Watching shifting grids while not knowing what
datums the grids where ment to be in is of no help.

Zitat

and what calibration method is supposed to be more accurate
(UTM vs. lat/lon), assuming that grid on paper is correctly printed and
labeled?

Provided the grids are printed correctly and the datum is known: For
simple navigational purposese both grids are absolutely sufficient.
Theoretical differences may be within 1-3 meter range.

It is not possible for me to judge the problem you see any deeper, as I
do not have the maps in hand. I have explained know what can be explained
about that matter. I confess, there may be half a study of cartography required to
understand the problematic relaions of datums/grid/possible errors.

Possible errors can be:
- wrong datum
- grids printed/located wrong
- error in calibration
- paper map stretched/distorted from aging/use
....and some more...

Teilen