OpenCPN Version 1.3.5 Beta Technical

[idpnd]

Originally Posted by PeteCross

Have you tried similar tests with OpenCPN?
Is there any way to ascertain which program is the most accurate?

I am not wanting to start a war with SeaClear, but I think the observation is interesting. Is there any other way to check this?

That is certainly an interesting and important comparison to be made!

Either the coordinates for a known location would have to be checked live, or you identify (mark) a position from the GPS stream on the map and then compare against the plotted position.

[Viking Sailor]

Pete,

How much position error (ft/m) are you seeing?

Paul

[PeteCross]

Originally Posted by Viking Sailor

Pete,

How much position error (ft/m) are you seeing?

Paul

Paul,

About 30 meters, which, if combined with normal GPS inaccuracy, is significant, I think. This difference might be explained some other way yet.
Its only my first observation, so there certainly is room for further investigation.

/Pete

[Viking Sailor]

Pete,

How fast is the boat going (kts or m/s)?

Is the error along the boats course?

Thanks,

Paul

[ata382]

Re position error:
i see same error with stored waypoints, ca 30 m offset with BSB maps in OpenCPN.
OK with CM93 charts in OpenCPN. Also OK with same BSB-maps in Seaclear.
Official charts for Sweden in higher latitudes are Transverse Mercator, not Mercator, but I think all CM93 charts are Mercator.
Could be a clue.


/Janne

[Viking Sailor]

Janne,

What is the direction of the offset?

Thanks,

Paul

[PeteCross]

Originally Posted by Viking Sailor

Pete,

How fast is the boat going (kts or m/s)?

Is the error along the boats course?

Thanks,

Paul

Paul,

The boat is doing about 7 kts.
The difference is mostly east/west, as far as I can see.
It might be relevant that this is found in Swedish waters outside Stockholm.
Can you not try experimenting from where you have your charts?
Its easier to compare if both programs leave a track on the chart.
Franson GPSGATE allows recording and playback of NMEA data, so this data can be used for comparison.

/Pete.

[PeteCross]

Originally Posted by ata382

Re position error:
i see same error with stored waypoints, ca 30 m offset with BSB maps in OpenCPN.
OK with CM93 charts in OpenCPN. Also OK with same BSB-maps in Seaclear.
Official charts for Sweden in higher latitudes are Transverse Mercator, not Mercator, but I think all CM93 charts are Mercator.
Could be a clue.


/Janne

I see from the Sandhamn.kap chart that it is indeed Transverse Mercator projection.
I hope Dave can give some guidance here, as my knowledge in this area is limited, to say the least and he does have a copy of just this chart, from a previous problem I had.

/Pete

[idpnd]

OpenCPN only supports mercator charts, so there you go..

This low level of distortion, combined with the conformal property which it inherits from the Mercator projection, make the transverse Mercator projection ideal for mapping regions with a narrow longitudinal extent, such as Chile. The European Environment Agency recommends its usage for conformal pan-European mapping at scales larger than 1:500,000

Although this might be clever up north

[bg9208]

Just downloaded OpenCPN - looks brilliant and can't wait to use it. Only one problem for me. When I click on object attributes there are a number of codes which are self evident and a number which I don't understand such as:- $scode,
_extgn,sorind, exclit, catrea, sigper etc. Is there available a dictionary explaining these? Also Drval1 and Drval2 are obviously drying values but relative to what MLW, MHW or chart datum?
Help would be appreciated.

[GPS-Marco]

Originally Posted by idpnd

OpenCPN only supports mercator charts, so there you go..

Even if OCPN display correctly only mercator, there is the possibility to avoid arrors at the centrer of the screen assuring that always the lat long of the centre of the screen is exactly on the right lat lon of the chart independently by the projection. This would minimize errors on the displayed chart.

Another issue is the georeferencing algoritms that in OCPN calculates the poly: if those are not right the chart is not properly aligned. It should be interesting to see the opencpn.log when the stocholm chart is loaded on OCPN screen.

Ciao, Marco.

[ata382]

Re position error:

Please see attached screenshots from OpenCPN and Seaclear with AIS messages in Oeckeroe harbour in Sweden. Same charts (Transverse Mercator).

I put the cursor at the end of the narrow pier in the harbour and got following positions:

SeaClear:
Lat. 57 42.339 N
Lng.11 39.596 E
OpenCPN:
Lat. 57 42.315 N
Lng.11 39.609 E


Diff in Latidude 0.024 minutes -> 0.024 * 1852 = 44 meter
Diff in Longitude 0.013 minutes -> at Latitude 57 degrees 1 minut of Longitude corresponds to ca. 1000 meter -> 0.013 * 1000 = 13 meter

Perhaps some tweaking ofthe georeferencing algorithms as proposed by GPS-Marco could improve the situation?


/Janne

[PeteCross]

Originally Posted by GPS-Marco

Even if OCPN display correctly only mercator, there is the possibility to avoid arrors at the centrer of the screen assuring that always the lat long of the centre of the screen is exactly on the right lat lon of the chart independently by the projection. This would minimize errors on the displayed chart.

Another issue is the georeferencing algoritms that in OCPN calculates the poly: if those are not right the chart is not properly aligned. It should be interesting to see the opencpn.log when the stocholm chart is loaded on OCPN screen.

Ciao, Marco.

Hi Marco,

There is no extra information in the log file when the Sandhamn chart is loaded to screen, apart from this below...

7:51:43 PM: Loading chart db version: V015
7:51:43 PM: Chart directory C:\CM93ed2_2009
7:51:43 PM: Chart directory C:\GPS\Charts
7:51:43 PM: GPS Watchdog Timeout is: 6 sec.
7:51:44 PM: Initializing Chart C:\GPS\Charts\6144 Sandhamn (1999).kap

Doesn't look too informative, does it?

/Pete

[PeteCross]

Originally Posted by ata382

Re position error:

Please see attached screenshots from OpenCPN and Seaclear with AIS messages in Oeckeroe harbour in Sweden. Same charts (Transverse Mercator).

I put the cursor at the end of the narrow pier in the harbour and got following positions:

SeaClear:
Lat. 57 42.339 N
Lng.11 39.596 E
OpenCPN:
Lat. 57 42.315 N
Lng.11 39.609 E


Diff in Latidude 0.024 minutes -> 0.024 * 1852 = 44 meter
Diff in Longitude 0.013 minutes -> at Latitude 57 degrees 1 minut of Longitude corresponds to ca. 1000 meter -> 0.013 * 1000 = 13 meter

Perhaps some tweaking ofthe georeferencing algorithms as proposed by GPS-Marco could improve the situation?


/Janne

Janne,

This seems a good method to calculate differences here. Much less complicated than my method and more accurate, but that was just the way I first saw the problem.

Anyhow, I too have checked using your method, but up here on the east coast at Sandhamn. I used the base of a bouy on the chart outside Sandhamn. These are my findings:

SeaClear:
Lat. 59 15.914 N
Lng.18 58.163 E
OpenCPN:
Lat. 59 15.907 N
Lng.18 58.224 E

Diff in Latidude 0.007 minutes -> 0.007 * 1852 = 13 meters
Diff in Longitude 0.061 minutes -> at Latitude 58 degrees 1 minute of Longitude corresponds to ca. 1000 meter -> 0.061 * 1000 = 61 meters

Quite a difference here, more than I had estimated from my previous attempt!

Thanks for the help, so far...

/Pete

[PeteCross]

Position error:

At this point, it would be good, if there were some official reference points at known locations around here, so we could verify our results.
Perhaps they exist, anyone?...

/Pete.