Charts II: NGA - 2700 Charts

[HappySeagull]

Originally Posted by bcn

......etc.......
There is still the question how to "release" a chart, declaring it qualified to be used.
Recri's scan is very helpful to get rid quickly of obvious errors but this does not prevent from having to inspect the charts one by one checking against the lat/lon grid.

Perhaps a scheme for chart status of
- first check for release done
- second check for release done/locked
might be an approach.
So release after inspection by two (different) persons.

A howto for inspections and things to look for? What are the parameters to declare a chart "done for release"?

Hubert

All they need is a BIG heading NOT FOR NAVIGATION.
Anybody that sails in the dark with just these is "winging it".
That said,anything "Coastal and Harbour" (<1:150000?) should get closer scrutiny.
...for me,if they show up in the location they are supposed to,with the correct Metres /Fathoms and geographic datum noted they are good enough for me to have handy when I'm thereabouts.. ..Oh!but we sure want local users to report back!How do they match with reality?with their tested and official charts? and I don't mean just Google Earth either,good as it is.
...and then there's the NTMS...

[HappySeagull]

a watermark
it's brutal,right there on the water but nobody can then be fooled...
-they won't show up on ebay,
-a crewman won't be slipping these into the ship's computer "as a favour"
-nobody on watch can refer to these without a pause.
-but they are still useful

[cagney]

Originally Posted by recri

Even transverse mercators are supposed to be rectangular charts, no?
.....

No, actually they're not quite rectangular.
They may appear to be, with the large scale charts that we normally calibrate.
Carefully checking the pixel-counts reveals that they are not.
To see this in a more extreme view check these pictures.
NGA38280




Or, check the very good NZ chart images (tif).

Of course, as you say,.... there are warped Tmercs as well in our little collection.

Thomas

[recri]

Originally Posted by cagney

No, actually they're not quite rectangular.
They may appear to be, with the large scale charts that we normally calibrate.

...

Thomas

Oh, that clears several things up. I'll see if I can compute the expected trapezoid. But in the meantime, I'll just leave the TM's out of the summaries. Leading to:

44285 corrected NE lng
43058 corrected to transverse mercator
43143 corrected to transverse mercator
43123 corrected to transverse mercator
43303 corrected to transverse mercator

[bcn]

Thomas,

a question:
which projections we can deal with?
I found some Gauss or Lambert Conformal Projections.

Are those supposed to be transformed into a valid .kap file?


Hubert

[cagney]

Originally Posted by bcn

Thomas,

a question:
which projections we can deal with?
I found some Gauss or Lambert Conformal Projections.

Are those supposed to be transformed into a valid .kap file?


Hubert

Just looking at OpenCPN, the situation is like this.
Lamberts Conformal Conical can't be used.
Gauss is basically another name for Transverse Mercator, and can be handled, to a point. Se this tracker entry:FS#641 : Tmerc issues
Most charts that OpenCPN can't handle are used for Polar Navigation. In other words, charts for very high latitude.
An interesting exception is chart 71343. 1:50000 in Gnomonic projection.
Gnomonic is used for Great Circle Charts, but was also a BA standard for large scale plans in bygone days. Nowadays Tmerc rules in that area.
1:50000 is small enough to make Mercator a bad approximation. This is apparent in the created kap. I have put a not about this in the comment field.
Thomas

[bcn]

Originally Posted by HappySeagull

All they need is a BIG heading NOT FOR NAVIGATION.
Anybody that sails in the dark with just these is "winging it".
That said,anything "Coastal and Harbour" (<1:150000?) should get closer scrutiny.
...for me,if they show up in the location they are supposed to,with the correct Metres /Fathoms and geographic datum noted they are good enough for me to have handy when I'm thereabouts.. ..Oh!but we sure want local users to report back!How do they match with reality?with their tested and official charts? and I don't mean just Google Earth either,good as it is.
...and then there's the NTMS...

Perhaps there are two points in this context:

- quality of the charts (basis and our work to make them usable)
- legal questions of how to use those


Part one: as in every (paper) chart you get you have to look for Notices To Mariners and you have to use your brain (I know you are clearly aware of this of course).
The NGA chart for my home harbor is 15 years old so it lacks to show the latest (500m) extension of the port. But on the other side it is spot on right now with all the rest - different to the CM93 for the same place (aprox. 15m south offset).
What we can do is assure the best fit from the material we are working with.
That is what I ment with quality. Making the optimal from it. Dot.

NTMS or updating is an issue - on the other hand: who is really updating his paper charts??
For example for the Barcelona harbor paper chart you get a sticker 5x5cm to update - because it changed a lot. (the basic chart is the one we are working with). The main river entrance went some half a mile south for instance.
Updating scanned charts would be a challenge with any doubt.
Local knowledge is another point to increase quality for sure.
But is this the project? Or is it sufficient to make the NGAs "as is" working for OpenCPN? In this moment. This is already a challenge - an it is looking good.

Part two: putting a note on the chart as big as possible will not save you in the Land of of the Free (USA). You are just at the mercy of some judge who will decide. That is valid for Furuno, Garmin, Jeppesen(Boeing) and all the others as well.
I don't care (at this moment) about the lawyer that wants to bite your ankle.

Hubert

- thanks!!

[recri]

Originally Posted by bcn

Thomas,

a question:
which projections we can deal with?
I found some Gauss or Lambert Conformal Projections.

Are those supposed to be transformed into a valid .kap file?


Hubert

According to wikipedia: The ellipsoidal form of the transverse Mercator projection was developed by Carl Friedrich Gauss in 1825[5] and further analysed by Johann Heinrich Louis Krüger in 1912.[6] The projection is known by several names: Gauss Conformal or Gauss-Krüger in Europe; the transverse Mercator in the US; or Gauss-Krüger transverse Mercator generally.

[dacust]

Originally Posted by bcn

...
Part two: putting a note on the chart as big as possible will not save you in the Land of of the Free (USA). You are just at the mercy of some judge who will decide. That is valid for Furuno, Garmin, Jeppesen(Boeing) and all the others as well.
I don't care (at this moment) about the lawyer that wants to bite your ankle.

You don't know that putting the notice on the chart will not help. You only know that putting a note on the chart is no guarantee you will be saved. But it might save you. A judge might decide that was enough. But if the warning is not on there you are guaranteed the warning will not save you, because you didn't even try.

I'm afraid that's how the system works. You have to try to put every warning out there you can. If you are lucky, one of them was the right one.

-dan

[HappySeagull]

Hi,Hubert...oh,the Law... I never thought about that!and care less,really.No,it's the poor b***ers that might buy these packaged up with opencpn for 20$ on Ebay that need know......or when they pop up on someone's screen, all-unknowing ,quilted in with "real"ones at an awkward moment-that- I worry about that.
Not because they are wildly bad or corrupt,or an attack on copyright -it's just that it's an avoidable confusion . I have too many aboard my haywire stringbag as it is...

much as I would hate to deface them,a Big and plain and tastefully faded watermark would not wreck their use at all?....Consider- if the NGA had done this(with sensitivity),wouldn't you still be happy to calibrate 'em?

PS: none of the above means I'm not keen to do a good accurate job of em ...and I even think their merits as "accurate" extends beyond boats.ie: someone in the Seychelles will appreciate these someday...you just never know.

[recri]

I just went through the "Not-doable in phase 1 - Composed of non-rectangular charts" category for Region 2 and corrected some charts to "corner cutouts not visible", "90 degree skew", or "partially doable, non-rectangular insets".

There were a few which I left as "composed of non-rectangular charts" even though their corner cutouts were not visible.

[bcn]

Originally Posted by recri

I just went through the "Not-doable in phase 1 - Composed of non-rectangular charts" category for Region 2 and corrected some charts to "corner cutouts not visible", "90 degree skew", or "partially doable, non-rectangular insets".

There were a few which I left as "composed of non-rectangular charts" even though their corner cutouts were not visible.

Hi Recri,

if there is a chart analyzed and the data submitted and it shows up that the corner cutouts are not visible it should be marked as such. That's the only way that Nohal can correct this. If it is marked as "partially doable" or "composed" and it is missing PLY's the .kap will just not be generated due to missing information - and that stays that way until somebody does mark it as one of the "corner cutout" category.

- 90 degree skew"
- "corner cutouts"

are works stoppers and should be marked as such in a first step.

Hubert

[nohal]

Hoy gang,
I've finally found a wi-fi reliable enough to do some bigger operations, so the corners are now being extended for the 137 charts still not doable...
Please allow about 2 hours for the batch to complete (as I write this, region 2 is processing) and clear the browser cache before doing the calibrations.
I hope I got it right, if not, let me know...
I have to say our calibrations are usually better than the Maptech produced charts in the area I'm now sailing, thanks to everybody involved!

Cheers

Pavel

[wdkester]

Everybody doing "new corner" charts remember to clear cache, I didn't.

[z.stadler]

Also, everybody who handles the "Corner cut-outs extended", please remember to change the status to one of the other options. Hopefully, it's "Doable in phase 1"

Clearly, if any corner is still invisible, set it to "... corners are not visible in cutouts"