Charts II: NGA - 2700 Charts

[dacust]

Originally Posted by HappySeagull

..it's really not that bad on the chart...
http://www.cruisersforum.com/forums/...tml#post793400

and
Attachment 32867

you can see the soundings through em .Transparent,right?I don't know what kind of nuisance it'd be to batch the jpegs with one,but easier than programming??probably bigger than this is better ....anyways,if we're all going to jail,we want Dave on the outside to bail us out...

This works perfectly. It has to be done on the chart, these charts can be used in programs other than OpenCPN.

I'm sure there are many ways to do a watermark, but ImageMagick can definitely do it.

-dan

[r.fairman]

Here is the warning page from Navionics, Remember they are supplying the software and the charts as we will be doing

[idpnd]

Why don't you just copy the exact wording from the NGA chart source

"View NGA's approximately 2,700 nautical charts, which provide global chart coverage between 84 North latitude and 81 South latitude. The NGA charts are NOT corrected for Notice to Mariners, but a link to the latest corrections corrections is given for each chart.

Use these charts as a ready reference or planning tool. Navigate with the official printed or digital charts"

"Note: Use official, full scale nautical charts for real-world navigation. These are available from authorized nautical chart sales agents. Screen captures of the charts available here do NOT fulfill chart carriage requirements for regulated commercial vessels under Titles 33 and 46 of the Code of Federal Regulations."

[z.stadler]

Originally Posted by idpnd

Why don't you just copy the exact wording from the NGA chart source

"View NGA's approximately 2,700 nautical charts, which provide global chart coverage between 84 North latitude and 81 South latitude. The NGA charts are NOT corrected for Notice to Mariners, but a link to the latest corrections corrections is given for each chart.

Use these charts as a ready reference or planning tool. Navigate with the official printed or digital charts"

"Note: Use official, full scale nautical charts for real-world navigation. These are available from authorized nautical chart sales agents. Screen captures of the charts available here do NOT fulfill chart carriage requirements for regulated commercial vessels under Titles 33 and 46 of the Code of Federal Regulations."

I think it is an excellent idea!

[dacust]

Originally Posted by r.fairman

Here is the warning page from Navionics, Remember they are supplying the software and the charts as we will be doing

OK, this may be totally nitpicking on your exact wording, but...

We are only providing charts, no software. The charts are not connected to OpenCPN in any way, except we are using the OpenCPN forum to do it in.

Right?

Originally Posted by idpnd

Why don't you just copy the exact wording from the NGA chart source

"View NGA's approximately 2,700 nautical charts, which provide global chart coverage between 84 North latitude and 81 South latitude. The NGA charts are NOT corrected for Notice to Mariners, but a link to the latest corrections corrections is given for each chart.

Use these charts as a ready reference or planning tool. Navigate with the official printed or digital charts"

"Note: Use official, full scale nautical charts for real-world navigation. These are available from authorized nautical chart sales agents. Screen captures of the charts available here do NOT fulfill chart carriage requirements for regulated commercial vessels under Titles 33 and 46 of the Code of Federal Regulations."

Dang. What a stupid idea. I wish I had thought of it first, then it would have been brilliant.

OK, OK, yep, I think their disclaimer might be enough.

-dan

[lhsmith]

For those who haven't been around to follow this long discussion, would someone restate the status of the 2700 charts, and if available in some form, where to download them.
Thanks,

[bcn]

Dear Larry,

the process to make the charts available for a broader use consists in several steps:

1. Downloading and stitching the jpg tiles (sub-pictures) from the NGA site.
This is completely done.

2. Classifying the charts - done

3. Georeferencing the charts, recovering the datum information, correct the charts (skewing, warping) and creating the kap-files.

You will find the actual status here:
NGA Charts project stats

Up to the moment we have some 2240 charts processed.
Some of those are having insets, that are not referenced nor processed.
You can test those kap-files if you register.

For the remaining we have to set up additional procedures to be able to generate the rest of the charts (non rectangular, charts with indents or extensions, chart sets on a single chart sheet, insets).

Please take into account what idpnd cites from the NGA site about the intended use of the charts (#993, 4 entries above).
"Notices to mariners" for corrections of the charts are not applied.
Some of the charts are quite old.

I'm saying "test them" as we do not consider the work as a release, ready to go.
We have not started yet:
4. Releasing and bundling charts for download.
5. Establish a maintenance cycle for the charts.

You are invited to test and to participate of course. We will be happy about your feedback. If you want to gain some more insight how the charts have been processed, there is a "Calibration Manual" on the site as well.

Hubert

[lhsmith]

Hubert -- thank you for your kind and informative update. I've duplicated my userid here at OpenCPN and will help in whatever way I can. Fair winds, Larry.

[richardhc]

A very big word of caution here.

The latitude and longitude shown on at least some of these charts are NOT WGS84 compliant even where the chart says they are. I downloaded 82646, calibrated it in MapCal, and compared it in SeaClear with the French chart from the same area. As part of my hydrographic surveys for the cruising guide to New Caledonia I had existing GPS waypoints for key geographic features on the French Chart. Plus I have my recorded vessel track of the same area. The French chart is correct. The US chart is significantly off.

You can't use the latitude and longitudes printed on these charts to calibrate them. The US chart actually notes that the latitudes and longitudes on the chart may differ from other charts of the area! You can see that right away by noting where the 164 meridian crosses the land on the following images.

Here is a screen shot of the French chart with my vessel track in red, a plotted entry route from the cruising guide to new caledonia and my gps waypoints as red squares. Note the position of the 164 meridian on the land.



Now here is a screen print of the US chart 82646 calibrated in MapCal using the latitudes and longitudes on the chart with the same track, route and waypoints - note the position of the 164 meridian on the land. And where my actual gps recorded vessel track crosses the land.



You can see that the blue route plotted on both images begins right on the 164 meridian, demonstrating the calibration is correct for the US chart even though the lat/long grid does not correspond with the real world. There is zero way for you to know which chart is correct without having actual GPS positions to verify the calibration. So I don't see how these free charts can be calibrated. Using Google Earth is risky because their grid does not always match reality either. It is pretty good in parts of New Caledonia now but just a few months ago it was incorrect. Again, without actual ground reference points you just can't tell.

Originally Posted by cagney

Now, through this thread -http://www.cruisersforum.com/forums/...tml#post718487, I have found out that there are ~2700 chart pictures, world wide available on the net. The pictures are in tiles stitched together, to one flash picture, something like a new version of the old hdr files.
Let us together work on this. First the picture wizards ... on how to make these pictures small enough and good enough ...please.

Thanks to NGA for the web site. A worldwide coverage of freely available charts is now within reach. If only BA could take the hint from NOAA, NGA, LINZ and the Brazilian Hydrographer........

Thomas

[bcn]

Hi Richard,

thanks for your comments that are very welcome.
We take this as a part of testing and quality assurance and would be happy if more user would report back.

One thing we recommend for validating of the generated kap files is to test against other charts (yes, they can be wrong as well) and if there is local knowledge and data as in your case even better.
Overlapping charts give valuable checkpoints as well.

If you have a look at the 82649 and have the outlines of the 82646 (kap) shown one can see the deviation comparing the printed (on the 82649) and the generated outlines.

Has somebody the 82646 paper chart? It should show the same error.
If this is the case that should be reported to NGA.

Hubert

[cagney]

Originally Posted by richardhc

.......

The French chart is correct. The US chart is significantly off.
.....
.

First, I'm pretty sure that NGA has not done any independent surveys of New Caledonia, and are basing their charts exclusively on French charts.
A general rule for charts is to always, if possible, use charts from the country that is responsible for the surveys. One reason for this is that you often find more detailed charts, not just aimed at ships entering big ports. Another reason is that new surveys and updates are first published by such a country.
We have earlier had this illustrated by an example from the Swedish coast in Gulf of Bothnia.
Exactly the same applies here.
The NGA 82646 2:nd ed was issued 1995 and based on French surveys "to 1987". Hence, we don't really know when the actual area in the pictures was surveyed.
The SHOM chart 7320 3:rd ed was issued 2011, and is supposedly based on later survey.
I'm fairly sure that older French charts would agree with the NGA chart.

So how can both these charts be to WGS84? I think that we will find a lot of cases like this in the future. Chart agencies will be forced to revise earlier data, based on incomplete and older surveys, from pre gps days.
This is a good illustration to why one should not trust charts just because they are "WGS", specially together with a gps.
Note that the US chart is perfectly alright unless you use, and trust the gps, without crosschecking with other navigation methods, ...and don't use your #1 eyeballs,.... to enter Baie de Banaré.

Thomas

[Tia Bu]

[QUOTE=cagney;807842 I think that we will find a lot of cases like this in the future. Chart agencies will be forced to revise earlier data, based on incomplete and older surveys, from pre gps days.
This is a good illustration to why one should not trust charts just because they are "WGS", specially together with a gps.
Note that the US chart is perfectly alright unless you use, and trust the gps, without crosschecking with other navigation methods, ...and don't use your #1 eyeballs,.... to enter Baie de Banaré.

Thomas[/QUOTE]

I'm not sure how much or how quickly the chart agencies will be "forced" to revise data. They are constrained by budgets, manpower limitations, government regulations, etc. NGA's mission is basically to support the military, I think. Would this kind of change really be relevant to the U.S. military?

On the other hand, any of us could take the two graphics above and come up with a pretty good estimate of a lat/lon correction that would make this chart far more accurate and useful for cruisers. Or- at the least, we could post a prominent note on the chart to warn folks of the observed error.

This example is a good argument for some sort of wiki revision mechanism in this project, IMHO. From what I can see, the chart is around a half-mile off, and the problem seems to be in the printed lat/lon grid on the NGA paper chart. The lat/lon error does not seem to have been addressed so far in Notices to Mariners.

This example also strongly argues Cagney's point: folks shouldn't be using GPS alone to navigate. A couple of years ago, I overheard a boat on the VHF navigating into the anchorage at Warderick Wells Cay in the Bahamas using his chart plotter. It was around 1:00 AM on a moonless night. He made it. That made me think maybe my navigation methods were just too old-fashioned. So a couple of weeks later, I tried the same nonsense at Bimini in the middle of the night. I ended up spending the rest of the night pounding on a sandbar. The GPS wasn't wrong, as far as I know. The chart was.

[richardhc]

Believe me, not all the French charts are reliably positioned for WGS84.
The new ones are generally OK but any French chart prior to (I believe) 1990 should be suspect.

Even if you start out with a chart that actually is WGS84 standard there are still potential errors from the scanning and when someone calibrates them.
I have, during the creation and maintenance of the Cruising Guide to New Caledonia , carefully calibrated all of the charts for New Caledonia (US, British and French) using real-world GPS measurements of landmarks and Seaclear's MapCal. (I made these ground-truth surveys anyway to calibrate the satellite and aerial images used in the cruising guide so I could overlay the depth/GPS coordinates on the images). I did the same for the US Charts (and satellite and aerial images) when creating the cruising guide to Vanuatu.

It would be very advantageous to have a "calibration score" on these charts. Interested participants of this forum or the suggested wiki could score the calibration by actually checking known geographic features on the chart with a GPS or by some other means.
Of course this assumes the participant follows some basic rules on doing a Ground Truth survey.

Suppose this quality check would work like this:
1= Checked in 3 locations with a GPS for specific landmarks in widely different parts of the chart and the chart was correct.
2= Checked in 1 location with a GPS for one specific landmark and the chart was correct.
3= Checked in 3 locations with Google Earth and the coordinates agree (the chart may be wrong and GE might be wrong but the chances of them both agreeing are small unless they are both correct - still it is only a third-rate score).
4= Checked against another electronic chart in 3 locations and the coordinates agree (in this case both could very well be wrong since electronic charts are based on the paper copy to begin with so a score of 4 means to use with caution but probably fine for large scale charts for open water route planning).
5= Not checked - use with caution.

There should be a second set of indicators to show if a particular KAP file has or has not been re-calibrated based on real world GPS measurements. The user would need to know that the particular KAP actually in his chart list was the recalibrated one and not one that was downloaded without calibration. This could best be done by renaming the KAP file by adding a C Before the chart number to indicate it has been Calibrated.
In that case the score would be 1C= checked with GPS in 3 separate areas on the chart and the chart was recalibrated.

[bcn]

Those who are having a look from time to time at the NGA charts statistics might have noticed a new category which is in place since yesterday:

"Additional PLY points defined"

Nohal added the capability to edit charts with indents or extrusions.

The process is as follows:
- Download KAP and GPX from the calibration website.
We recommend to keep the files you are working with in an extra directory, disabling all the rest of charts with exception of perhaps your reference set.
(Rebuilding the chart DB in OpenCPN might be needed to see the changes in the new version of the chart)

- Import the GPX to OpenCPN's route manager. It will follow the rectangular contour of the chart
- Zoom to the route for defining the extents of the chart
- Turn off quilting (F9)
- Do the insert waypoint/move waypoint modifications to the route to follow the desired shape of the chart
- Export the route (the name has to be <chart_nr>.gpx again)
- Send the route back to the backend ("Click here to submit the chart polygon as GPX." link next to each chart)


We tested it yesterday and are aware that there might be more elegant solutions but it works fine.

So we can go ahead to edit another group of charts.

The tool might give us the way to clip off texts, other symbols, or insets that are placed "on the water" - common thing for islands.
There is no operative set up for this kind of editing, so please stay tuned while editing the "Special PLY's needed" set of charts!

Hubert

[wdkester]

Hubert,
When I try to "Import the GPX" in IE, Chrome, and Firefox, I get a listing of the script. It doesn't execute. Crome gives the following:
This XML file does not appear to have any style information associated with it. The document tree is shown below.

<gpx xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://www.topografix.com/GPX/1/1" version="1.1" creator="GPX RubyGem 0.6 Copyright 2006-2009 Doug Fales -- http://gpx.rubyforge.org/" xsi:schemaLocation="http://www.topografix.com/GPX/1/1 http://www.topografix.com/GPX/1/1/gpx.xsd">
<metadata>
<name>21005.gpx</name>
<time>2011-10-04T22:00:18+02:00</time>
<bounds minlat="90.0" minlon="180.0" maxlat="-90.0" maxlon="-180.0"/>

</metadata>


<rte>
<name>21005</name>
<rtept lat="27.0001388888889" lon="-120.500944444444">
<time>2011-10-04T22:00:18+02:00</time>
<ele>0.0</ele>

</rtept>


<rtept lat="30.7501388888889" lon="-120.500944444444">
<time>2011-10-04T22:00:18+02:00</time>
<ele>0.0</ele>

</rtept>


<rtept lat="30.7501388888889" lon="-113.834277777444">
<time>2011-10-04T22:00:18+02:00</time>
<ele>0.0</ele>

</rtept>


<rtept lat="27.0001388888889" lon="-113.834277777444">
<time>2011-10-04T22:00:18+02:00</time>
<ele>0.0</ele>

</rtept>


<rtept lat="27.0001388888889" lon="-120.500944444444">
<time>2011-10-04T22:00:18+02:00</time>
<ele>0.0</ele>

</rtept>



</rte>



</gpx>

The "export route" script executes but I don't have anything to export yet.