Charts II: NGA - 2700 Charts

[Surveyor]

This is a team effort but I am closing in on 800 charts downloaded! I do not want the glory though, I want a full set of charts!

Considering the hours already spent and the certainty that hundreds more will be forthcoming, I better not hear a single complaint about the charts!

[wdkester]

General Subject - Chart QA

What I would like to discuss is the idea that there are at least two kinds of quality: Type 1. Did our process accurately reproduce the chart that DMA (the author) intended and Type 2. Does the DMA chart accurately portray the intended area. (ie how does it compare with a more accurate chart or with Google Earth) I think both are important, but I think, that initially, we should concentrate on Type 1.

Background 1 – I have scanned DMA charts of Belize. There is very little coindence between GE and the charts; so lots of Type 2 problems. However, I have cruised Belize using these charts. For tight places (like Porto Stuck) I made “Google Earth charts.

Background 2 – I have some scanned charts that were produced using recent technology. Where Google Earth has good data, the two compare nicely. In 20 minutes of checking , I have found 2 anomalies in the charts: 1. the text in the cruising guide had bad coordinates for a WP; 2. a WP symbol was placed in the wrong location on the map (the printed coordinates were correct). It’s hard to say who made the error, the author or the printer, but think these are largely Type 1 errors. The printer didn’t reproduce what the author intended

In checking the lat/lon lines of scanned charts I find “small” errors. Also sometimes the lines are not entirely straight. They have several pixels that are offset like this
-------______-___----.

Now we get to the QA part. First lets define lat/lon in DDD MM.1234, where number represents the decimal place.

Case 1: Using the DMA chart 28162 for Belize I zoom in until I get the jagged line effect, Opencpn gives me 15 40.0051 / 88 55.0127 for a junction in the lower left hand corner. The scale for this chart is 1:150000, the lat/lon increments are 10 min., and the line widths are about .01 min. (Jump in if I’ve made a math error) I think .01 min error would be 6.1 feet or less than 2 meters As far as I am concerned the lat on this chart is very close, less than a line width off. The long is getting close to GPS and line width error, but still, I think, acceptable.

Case 2: Using one of my Background 2 charts, I zoom in until I get the jagged line effect,; I move the cursor to a lat/lon junction; the chart indicates DDD MM.40/DDD MM.80 Opencpn reads DDDMM.3995/DDD MM.7985. The scale isn't given; the lat/lon increments are 0.10 min and line widths seem to be about .0015 min. To me this is quite acceptable tolerance for Type 1 errors . GPS error is larger than this.

To summarize, Type 1 errors should be less than .01 min. or the line width whichever is larger. Experience may dictate somewhat relaxed standards. Type 2 errors should wait for a few months. I can imagine a rating system where an AA rating meets the highest standards for Type 1 and 2 errors.

[dacust]

Originally Posted by HappySeagull

...part of my thinking per gpx is that the NTMS would show up on ALL your charts ,if wanted.

This,as a layer,withgpx marks using custom icons, is a great feature of O....Other apps will follow OpenCPN and read her GPX or wither and die in their proprietary formats(curse them!)

That said,anyone that wants to can extract any kap into an image and rebuild it,per freely available Notices to Mariners can do so right now.

So!-Those of you that've rebuilt a chart every time a new NTM comes out,raise your hands!

Using OpenCPN layers:

• NOTMs show on ALL charts in OpenCPN
• Do NOT show in any other software

Editing existing charts:

• NOTMs show on NGA charts in ALL software
• Do NOT show on other charts

These are almost 2 different projects.

One project is to maintain NOTM layers in OpenCPN.
The other is to apply NOTMs to the old outdated NGA charts we are currently processing.

I don't see it as a choice to do one or the other, but to do both.

Applying NOTMs to existing old charts, you take the date of the chart, and start applying. You know where to start. If a buoy moves, you can wipe out the old one and add the new one. The chart stays clean.

Applying NOTMs in a layer, how far back do you go? If a buoy has moved, the old one will still show on the screen, do you include another object like a red X on top of the old buoy? The chart can get really cluttered after a while. With charts of different dates, you may have a new layer buoy on top of an existing buoy, leaving the user wondering what may have changed - or with lettering on top of each other making it hard to read. I think the layers idea is really good, but I think in practice it may be hard to implement well.

-dan

EDIT: I should have added: Just because I see issues in trying to implement NOTMs in layers doesn't mean I am saying we should give up on the idea. It has too many advantages to simply abandon. But my gut feeling is that in the end it may just not be practical.

[HappySeagull]

Dacust,all true.I agree,I'd like to have the image corrected but for the work involved...and I haven't DONE a gpx NTM yet,so it's all pi in sky.(Homemade icons would be involved (for buoys,soundings etc..)or maybe just a note placed on screen....?)

[Viking Sailor]

As much as I like and plan on using OpenCPN as my charting software -- I don't think that using layers is the best way to go. Mainly because there will always be new NOTM's to be applied to these charts. Basically we need to crowd source this process in an on going way.

Looking at the numbers -- there are 2700 chart that need to be corrected. If we could get something like 100 "navigators" to do one or two charts a day the whole thing could be done in about a month.

There are two things that well be needed to make this possible. First, is to associate the NOTM's with the charts. Second, provide "How-To" text and video instructions detailing the processes need to apply the NOTM's to the charts.

[cagney]

Originally Posted by r.fairman

QA

The QA using GE works well as suggested .

Drop 4 waypoints in landmarks on your chart
export as GPX
Open using GE
Do they line up Y/N
If Y calibration is ok
If N recalibrate and QA

As an aside it would be nice to use the GE plugin for this. Can waypoints created in OpenCPN display in the plugin without having to close the plugin, open GE and import

An alternative QA would be to move the boat to known landmarks and compare its position in the GE plugin if that is possible

regards

Richard

Geo-referencing accuracy and QA is a question of getting chart plotters to display the original charts the way the cartographer intended. It is not a question of adjusting the NGA chart-datum to agree with GE. GE has it's own problems and errors.
Waypoints in the corner won't do anything more that the PLY polygon already does.
A manual QA would mean checking a number of lat/long crossing in the middle region of the chart, to make sure that OpenCPN's lat&long values agrees with the chart.

Thomas

[HappySeagull]

per:
"Applying NOTMs in a layer, .... If a buoy has moved, the old one will still show on the screen, do you include another object like a red X on top of the old buoy? ...etc- or with lettering on top of each other making it hard to read."

...but "layers" are just gpxfiles and fairly easily editable in openCPN when opened not as layers, but as marks...if a buoy needs moving or a new icon-move it/change it in the dialogues.
.. but like I said,it needs some tinkering to see how it'd work.It might be faster to just use imgkap and extract/edit the chart but I do dig seeing NOTS in all charts...and also, not messing with the originals much...

[r.fairman]

Ok i stand corrected
So for QA we simply cross check a number of chart grid crossing points in the central region of the chart with what the opencpn grid is displaying So if both grids line up all is ok

One other point some charts are Loran C. Does our www based calibration process take this into account


regards

Richard

[Surveyor]

ATTENTION DOWNLOADERS

Is anyone with ruby grabbing/stitching any of the remaining few charts? There are 16 left that are apparently python failures. If your using python your probably wasting your time. Your effort is appreciated though!

I have updated the failed list. Items with no comments still need to be grabbed and stitched using ruby. You can find the list here:
Failed Downloads

Please reply in this thread if you have intentions of downloading or already have downloaded some of these.

I have already committed to grabbing the leftovers in Region 5.

[HappySeagull]

Surveyor,I got 21530 and uploaded.

[Surveyor]

Good timing, I was just about to do that one next. I got the rest of the 5xxxxs.

[dacust]

Originally Posted by HappySeagull

per:
"Applying NOTMs in a layer, .... If a buoy has moved, the old one will still show on the screen, do you include another object like a red X on top of the old buoy? ...etc- or with lettering on top of each other making it hard to read."

...but "layers" are just gpxfiles and fairly easily editable in openCPN when opened not as layers, but as marks...if a buoy needs moving or a new icon-move it/change it in the dialogues.
.. but like I said,it needs some tinkering to see how it'd work.It might be faster to just use imgkap and extract/edit the chart but I do dig seeing NOTS in all charts...and also, not messing with the originals much...

The original charts are raster. The original buoy will not be an object you can move. It's part of the bitmap image. The NEW buoy you add WILL be an object. It it later moves again, then that NEW object can be moved. But the old original buoy will still be there in its original position.

-dan

[Surveyor]

YAY! The rest of the charts are stitched and I am uploading them now. There are of course several that are missing tiles and one that just wont cooperate although all of the tiles are there.

I am declaring this a milestone accomplished!

[Surveyor]

Nice job communicating guys!

Someone is uploading the same files I am.......thats some nice teamwork there cowboy

I totally expect that there will be 3 people calibrating the same charts too.

[HappySeagull]

Originally Posted by dacust

The original charts are raster. The original buoy will not be an object you can move. It's part of the bitmap image. The NEW buoy you add WILL be an object. It it later moves again, then that NEW object can be moved. But the old original buoy will still be there in its original position.

-dan

...and that's where (over the the old buoy) I'll use the "X" icon I'm making...and this "correction" gpx would show on the vector charts too (CM93's would need be position-calibrated properly with OpenCPN to have the X right on the old buoy in them)
Oh,there's problems in my thesis,for sure.long Pipelines or powerlines would be more troublesome as a big icon laid in pieces than just drawing on a charttiff...but still possible.