Charts

[Totobeloeil]

Originally Posted by BluegillBoy

3) what I'm seeing in the file is:
bin\mc2bsbh -l %MAP% Calibration\CHARTCAL.DIR
(an "-l" rather than an "-s")
am I in the wrong place or has that command changed?

I would try this :

 bin\mc2bsbh-poly -p -s %MAP% Calibration\CHARTCAL.DIR

'-l' is for 'get the list of charts contained into infile (CHARTCAL.DIR)'
'-s' is to 'proceed for one single chart included into infile''

4) is there anyway to get an actual .bsb file? that seems to be the key to some other nav programs.

You can get free NOAA raster charts here : Raster Navigational Charts: NOAA RNCs
Some BSB files are included.

Jean-Pascal

[BluegillBoy]

I'm working with KapGenAll which probably explains the "-l" list versus the "-s" single chart.

NOAA is great for the lakes and coast. But they don't offer raster charts for the waterways and canals - or the Illinois and Mississippi rivers. I'm trying to provide a full library of those in raster format.

The United States Army Corp of Engineers is responsible for those waterways, not NOAA. And the scanned charts they provide have much more detail and information that the ENCs.

Which is why I'm trying to find if there's a way to get the .HDR or .DIR headers in a .BSB formt which many of the other navigation programs need.

[Totobeloeil]

Originally Posted by BluegillBoy

I'm working with KapGenAll which probably explains the "-l" list versus the "-s" single chart.

mc2bsbh-poly is a part of KapGenAll . It replaces mc2bsbh. The '-l' option has nothing to do specifically with KapGenAll (the loop is inside the command file). That is why I propose to change the command file and run it.

I'm trying to find if there's a way to get the .HDR or .DIR headers in a .BSB formt which many of the other navigation programs need.

KapGenAll will generate all the headers you need and incorporate them to the kap files. I'm not sure to understand why you absolutely want to make bsb files.

Jean-Pascal

[GPS-Marco]

Originally Posted by BluegillBoy

I'm working with KapGenAll which probably explains the "-l" list versus the "-s" single chart.

NOAA is great for the lakes and coast. But they don't offer raster charts for the waterways and canals - or the Illinois and Mississippi rivers. I'm trying to provide a full library of those in raster format.

The United States Army Corp of Engineers is responsible for those waterways, not NOAA. And the scanned charts they provide have much more detail and information that the ENCs.

Which is why I'm trying to find if there's a way to get the .HDR or .DIR headers in a .BSB formt which many of the other navigation programs need.

Hello,

KapGenAll.cmd is a batch that calls KapGen.cmd for all charts

The real work for each chart is done in KapGen.cmd

At line 41 in KapGen.cmd batch version beta06 you will find the line with the -s switch that you have to change to use the poly version of mc2bsbh.

If many people will use the mc2bsbh-poly version and do not report problems, we can make the -poly version the standard mc2bsbh (in fact the -poly version without the -p switch works exactly as the standard non poly version...)

About the BSB file, this seems not needed for many applications we have tested with the charts generated by KapGen/mc2bsbh and for sure this is not needed by the free ones OCPN and SeaClear (that are our primary target applications)

Ciao, Marco.

[HamishB]

Originally Posted by BluegillBoy

The United States Army Corp of Engineers is responsible for those waterways, not NOAA. And the scanned charts they provide have much more detail and information that the ENCs.

are they georeferenced or just flat scans? i.e. can `gdalinfo` tell you lat/lon information about them?


Hamish

[HamishB]

Originally Posted by GPS-Marco

Anyway I've seen that when OCPN fails in georeferencing charts (skewed charts, non mercator charts, etc.), in these cases mc2bebh-poly reduces considerably the error. Better than nothing... Give it a try.

w.r.t. "better than nothing", well that's not quite correct anymore.

These days I am happily reprojecting aerial photos, LANDSAT imagery, etc. from arbitrary map projections into Mercator and automatically chopping away non-navigable chart-edges and then exporting them as GeoTiffs and a suitably formed header file. From there you can use tif2bsb if it is legal to do so in your country. If not just enjoy the GeoTiff in your favourite mapping software (e.g. QGIS).

The script uses GDAL and GRASS GIS tools to ensure that the georeferencing is exactly right with no tedious human digitization required. (this isn't really meant for registering maps yourself, this is more for preparing geodata already in georeferenced form)


even without that, it is very quick to use `gdalwarp` to reproject the image to Mercator before registering it. (even though that seems like double-handling to me)


regards,
Hamish

[BluegillBoy]

"are they georeferenced or just flat scans? i.e. can `gdalinfo` tell you lat/lon information about them?"

they have lat/lon and gridlines printed on them. the issue has been that they're not north-up, which chokes MapCal.

however... I've been having a lot of success with Marco's poly! even tho MapCal says the E/W and N/S are reversed, it does save the ChartCal.dir and KapGen (with poly integrated into it) produces charts.

i'm gereally getting comments from KapGen projecting errors of betwee 0.6 and 2.0 pixels. (PS - I LOVE that error projecting feature, Marco. When I start seeing errors of 140 pixels, it's an instant prod to go back into calibration and find my typo!)

I'm launching the boat this weekend, and will try out the charts. sitting at home, they don't always seem to match edge to edge exactly when I scroll them with the mouse. But I really need field trials onboard to tell for sure.

more updates on Monday - I'm goin' cruisin'!!

[BluegillBoy]

Launched this weekend in 34-degree temps, cruised her 60 miles to our marina using OpenCPN and the charts I created - wonderful.

The original problem is the the Illinois Waterway charts aren't north-up, they're strip charts oriented to show the greatest length in a landscape format.

Got the PDF charts from USACE, calibrated using at least 4 points with MapCalII, converted using the Poly version by modifying KapGenAll to use the poly version - and that did the trick and let me do a batch convert on the charts.

All seemed accurate and seamless while underway.

One issue still remains - OpenCPN doesn't project the bearing line well. Sometimes it's accurate for the orientation of the chart, but sometimes it's projected assuming north-up even though the chart isn't north up. And sometimes it switches back and forth even on the same chart.

I can live with the "boat" symbol pointing the wrong way, but that bearing line gets distracting. Is there a way to just turn it off?

The other wish I have is for quilting, rather than only showing one chart and the njumping to the next. Wrong thread for a wish list I know.

Anyway, if anyone needs the Illinois Waterway in raster format, let me know. Always glad to share.

[HamishB]

Originally Posted by BluegillBoy

One issue still remains - OpenCPN doesn't project the bearing line well. Sometimes it's accurate for the orientation of the chart, but sometimes it's projected assuming north-up even though the chart isn't north up. And sometimes it switches back and forth even on the same chart.

perhaps there is a atan() instead of atan2() in the code somewhere?

or maybe just that the polyfit blows up when you try and push it too far?

it is recommended to rotate the raw scan with a graphics program like the GIMP before doing any registration. You can drag a horizontal or vertical guide from the left or top edge of the image canvas to help with lining up the grid lines while using the rotation tool.


Hamish

[BluegillBoy]

I have rotated the charts to north-up before calibration. the problem with the strip charts that the Army Corp of Engineers makes is that when you rotate them, the descriptions, words, markings, everything are then rotated 90 degrees or even upside down. So they work fine once calibrated, but really aren't readable and therefore aren't usable.

So I'm still wishing that OpenCPN would properly project bearing lines even when charts aren't north-up.

Any help with that?

[HamishB]

Originally Posted by BluegillBoy

I have rotated the charts to north-up before calibration.

ok, probably nothing to do with the kapgen polyfit then.

the problem with the strip charts that the Army Corp of Engineers makes is that when you rotate them, the descriptions, words, markings, everything are then rotated 90 degrees or even upside down. So they work fine once calibrated, but really aren't readable and therefore aren't usable.

So I'm still wishing that OpenCPN would properly project bearing lines even when charts aren't north-up.

Any help with that?

my guess is that it is the (recently fixed) bug of switching calculations when the distance was shorter than a certain amount? does the bearing line fix itself when you move it say more than 1/4 mile away from the boat but break when very near to the boat?


the program has no idea what is printed on the chart image, so no way of knowing that the original was not north up, has words on it, etc. the only solution to that is to use vector based ENCs or a lot of time manually rotating features in the image editor.


Hamish

[idpnd]

Originally Posted by HamishB

the program has no idea what is printed on the chart image, so no way of knowing that the original was not north up, has words on it, etc. the only solution to that is to use vector based ENCs or a lot of time manually rotating features in the image editor.

That's a good point actually - chart rotation won't be particularly pleasing on raster charts..

[GPS-Marco]

Hello. Here is new version of New Zealand Charts Portfolio for you. Changes from V23 to V24 are about 10 maps:

NZ53, NZ522, NZ531, NZ5225, NZ5225_1, NZ5411_1,
NZ5411_2, NZ7621, NZ7621_1, NZ7622

Reason of change is that LINZ on March 24th has released new versions of above map images (either for reprinting old charts or for adding some Notice To Mariners).

If you want them, download the new ChartCal-V24 ( MCE66 - Software ) and rebuild with KapGen the above charts (first delete the relative tif files from /Img folder if you have alredy built those maps). Or just download the already built charts (Upgrade File) from the link at the same location.

Let me know. Ciao, Marco.

[brak]

Originally Posted by idpnd

That's a good point actually - chart rotation won't be particularly pleasing on raster charts..

I beg to differ. Raster charts look just fine rotated. Of course I should say that because we implement it, but aside from that - I find raster charts no less legible than my paper charts which I usually keep aligned with boat heading. It makes finding land features, navaids etc a lot easier, at the obvious expense of text being potentially upside down.

[BluegillBoy]

And of course depths being difficult to read.

but the big difference to me is on the river rather than lake/coast. the river charts are printed on long thin pages, always with the river running left & right. Plus the things most important to read are position, which is generally given as "downbound approaching the Santa Fe Bridge" or the Wilmington Ave. bridge, or the Acme Steel docks. ATONs got nothin' to do with it except keep you in the channel. it's impossible to tell an approaching tow where you are if you can't instantly read your chart.

that said, there's good reason for various chart and, as you point out, open water navigation takes a whole different approach. Which is probably why open water charts are almost always printed north-up.