openCPN chart geo-referencing issue

sebastiannr

sebastiannr

Well-Known Member
Joined
4 Apr 2010
Messages
186
Location
Liveaboard, currently in BVI's
Visit site
Hi there,

I have openCPN with CM93v2 vector charts and also Maptech BSB raster charts of the Med (2001). I have noticed that the two chart types aren't geo-referenced the same. For example, I plot a route leaving a AgNik marina (Crete) using the CM93 charts, when I then change to the corresponding BSB raster chart it will show my starting position as being on land. I assume that it's an issue with the geo-referencing of the raster charts, but any kind of uncertainty is obviously unacceptable.

I'm not in the Med at the moment (flying out to Crete on the 21st) so I cant check it by using my actual position and comparing that to the charts.

If anyone has any thoughts on what the issue is and how it can be resolved I would be grateful to hear them.

Seb.
 
I've been able to reproduce the exact same problem at this location but not in the few other places I have tried out. The positions plot much the same in these latter checks so it would appear the chart datums are similar if not the same. It's probably an isolated case but something to keep an eye on. It shows the value of relying on the Mk1 eyeball. By the way, the charts Maptech uses for this area a German ones, I guess originally from the war, updated as necessary since.

Sebastiannr, do you know if the latest Maptech charts are any better than the 2001 ones?

Cheers,

Mike
 
It really depends entirely, as already suggested, on the map datum in use on the various charts. No doubt they are all nominally WGS84, though. But it isn't that simple. What I THINK you are seeing is maps originally on survey based datums that have been transformed to WGS84. A survey based datum is irregular - basically, you have a network of survey points whose location has been computed as precisely as possible using trigonometry (this is pre-GPS stuff), and those locations are then ASSUMED to be correct. Now, in the UK, the OS has run an enormous programme to create accurate corrections between these survey points and WGS84. But this is unlikely to have happened in less intensively mapped parts of the world, so you will find places where the correction from the original survey based datum to modern ellipsodal datums is incorrect.

The good news is that errors should not exceed a few hundred metres. The bad news is that they may be up to a few hubdred metres!

GPS may well give you an absolute position to a few metres, but that doesn't help if the error in the position of hazards can be far larger. And in many parts of the world - even as close as the Eastern mediterranean - the charting errors may well be that big.
 
Thanks for the info - it's both interesting and disturbing in equal measure.

The raster charts are nice as an overlay for planning pilotage, but if they're potentially inaccurate to that degree it pretty much makes them useless and/or dangerous to use for navigation.

Can I assume that the CM93 charts wont have those inaccuracies?
 
I've found large variations between Navionics (vector) and CM93 (vector) charting software. By observation and triangulation I've found the CM93 charts consistently more positionally accurate.
In my case I'm using two entirely stand-alone systems, with separate receivers as well as charts.
There can be any number of reasons for the inaccuracies, from datum differences, through receiver calibration or obsolescent origination charts to simple error.
By far the worst errors were from software based on Spanish and Italian chart origination.

As the CM93 charts are ECDIS approved and virtually no others are, I'd always plump for them. As it is I have a profound suspicion of all positioning devices and prefer to have at least 3 in agreement (if you include vision as a positioning device).
 
sebastiannr said:
Thanks for the info - it's both interesting and disturbing in equal measure.

The raster charts are nice as an overlay for planning pilotage, but if they're potentially inaccurate to that degree it pretty much makes them useless and/or dangerous to use for navigation.

Can I assume that the CM93 charts wont have those inaccuracies?
Click to expand...

No, you can't. The CM93 charts will be based on the best mapping that was available to the compilers, but that is precisely as good as the original survey - and you have found evidence of the quality of that survey.

In MOST parts of the world, it isn't safe to assume that charts are as accurate as a GPS. Only in relatively wealthy countries is there the funding available to ensure that charts are as accurate as (for example) Admiralty Charts in UK waters. In most places, the fundamental survey will be old, often done by colonial powers in pre-GPS days. The technology used means that there are intrinsic errors in the hundred metre range, and short of a complete re-calculation of the survey grid using new GPS locations, the charts will remain inaccurate. Such recalculation is a major effort, as it affects all sorts of industries you wouldn't imagine - the entire building and civil engineering sector, as well as marine interests, and others. The UK has done that - but as recently as 7 or 8 years ago, it was still an ongoing process (charts were mostly less affected and were done as part of the routine chart re-issuing).

The good news is that the relative accuracy of the charts will be fine - so you need to use old-fashioned methods of navigation, such as bearings, lines of sight and distance off. If using GPS in areas you are doubtful of, allow a few hundred metres safety margin!
 
You must log in or register to reply here.

Share this page

Top