Charts II: NGA - 2700 Charts

[recri]

NGA 2700 chart summary for Fri Sep 23 18:10:39 GMT 2011
processed: 1382 charts
evaluated: 1266 charts
pxhoriz: mean 1.00000 stddev 0.00017
>5*sd: 22531 22533 29325 24215 83039 24311
>4*sd: 81427 81616 55001 93032 36005 21182 83495
>3*sd: 82564 53091 56190 81127 42740 44187 54085 400 51150
pxverti: mean 0.99999 stddev 0.00017
>5*sd: 91407 29325 83039 82025 83105
>4*sd: 83495
>3*sd: 53290 56190 22533 53287 24319 53266 43340 81796 18766 14003 83010 26001 21182 61190 92200
pxsquare: mean 0.99999 stddev 0.00018
>4*sd: 83392 82025 83020
>3*sd: 29325 63361 82304 63354 81016 24215 24381
ratio: mean 1.00492 stddev 0.00179
censored: skipped 91 charts: 11 62 70 105 106 121 124 125 126 145 301 520 521 522 524 526 532 622 623 701 707 800 21141 21161 21621 21641 22100 22314 22404 22440 22528 22541 22542 22549 23051 23121 24152 24154 24159 24444 24511 26125 29282 29322 29324 37242 42036 42046 54289 54303 57022 61003 63015 71009 71012 71018 71036 71140 73034 74024 76201 81145 81303 81329 81511 81563 81565 81576 81729 81737 81791 82384 82534 83052 83218 83555 91020 91030 91060 91096 91314 91326 92025 92030 92033 92040 92050 92140 92190 92560 96036
notmercator: skipped 25 charts: 14221 17421 18419 28164 29323 43064 43263 43344 43363 43371 43377 43378 43386 44192 44310 54362 54364 54368 55064 56191 76171 76181 81327 82071 91286
-----
Processed all charts with status:
+ 'Rectangular, doable'
+ 'Rectangular, partly doable - rectangular insets'
+ 'Rectangular, partly doable - non-rectangular insets'
+ 'Rectangular, not doable in Phase 1 because of special PLY's needed'
+ 'Chart done, checked and locked'
'processed' is the number of charts processed
'evaluated' is the number of processed charts that didn't get skipped
'pxhoriz' compares the lengths of the N and S edges, which should be equal
'pxverti' compares the lengths of the E and W edges, which should be equal
'pxsquare' compares the lengths of the NE-SW and NW-SE diagonals, which should be equal
'ratio' compares the aspect ratios of the REF point rectangle and the projected corner
+ coordinates, which should be equal
All those 'shoulds' are assuming a rectangular chart area in mercator projection.
'>n*sd' charts are greater than n standard deviations from the mean
'>5*sd' should be none
'>4*sd' should be none
'>3*sd' should be 3/1000 of the charts processed
All those 'shoulds' are assuming a normal distribution.
'censored' are outliers previously checked and looked 'okay',
+ there may still be a problem with data entry, or the chart may just be crooked.
'notmercator' are charts that aren't mercator or unknown

[HappySeagull]

Originally Posted by recri

They get picked up on pxhoriz, pxverti, or pxsquare because the charts aren't as rectangular as most charts are, either they were drawn crooked to begin with or the chart paper got shrank or stretched a bit somewhere.

They get picked up on ratio because the chart isn't the same rectangle as the projected corner coordinates, the corner coordinates and the mercator projection specify an exact rectangle shape.

The datum or lack should have no effect, a purely imaginary chart could pass or fail the same tests.

...which encourages me to ask if there's a way to describe these as angles of twist or tilt? or just unequal sided...

[recri]

Originally Posted by HappySeagull

...which encourages me to ask if there's a way to describe these as angles of twist or tilt? or just unequal sided...

The pxhoriz and pxverti simply complain about opposite edges diverging from equal lengths. pxhoriz = 1 and pxverti = 1 implies a rectangle or a regular parallelogram.

The pxsquare complain about oppsosite diagonals diverging from equal lengths. pxsquare = 1 implies a rectangle.

All of these are computed ignoring rotational errors. They do not assume that the horizontal or vertical edges are aligned with the image coordinates.

I'd say most of what we're finding, once we sort out the bad coordinates and wrong corners, are charts that aren't square to the precision of the scanner. They were square enough until you put them under the microscope.

The amazing thing, to me, is that the charts that pass the test are microscopically square.

[HappySeagull]

recri:
ah!Diagonals!a good way to test, alright!I was thinking that the numbers looked like sins or tangents but I see better now.Thanks

[Doek]

I see that there are over 5000 charts in the NGA library. There seem to be many charts missing from the JPEGS we have. Will we get the other 2300+ charts eventually? What exactly does the 2700 charts cover?

[recri]

Originally Posted by quackedo

I see that there are over 5000 charts in the NGA library. There seem to be many charts missing from the JPEGS we have. Will we get the other 2300+ charts eventually? What exactly does the 2700 charts cover?

Interesting, where'd you get the number?

I count 2773 numbered charts from scrounging through the image directories at Index of /NGAViewer, I expect you can get over 5000 if you split out the insets and the charts with multiple plans on the sheet as separate charts.

[cagney]

Originally Posted by quackedo

I see that there are over 5000 charts in the NGA library. There seem to be many charts missing from the JPEGS we have. Will we get the other 2300+ charts eventually? What exactly does the 2700 charts cover?

Go to this page Index of /NGAViewer, scroll down, and download the pdf catalogs nga1 .. nga 10. There you will find the exact coverage and also the nga charts that are not available to this project, as they are crossed out in red.

Thomas

[paulr]

Originally Posted by recri

NGA 2700 chart summary for Fri Sep 23 18:10:39 GMT 2011

Region 8 checked as follows:

81016 - input is good (very small scale chart, old)
82304 - not doable in ph 1, requires special PLYs, but the data input is good
82025 - unknown projection, input is good (old chart)
83020 - input is good (old chart)

83105 - all good, but, in absence of adequate edge marking, SW and NE corners have been manually calculated (no idea how or to what accuracy).
83495 - all good, but, in absence of adequate edge marking, SW and NE corners have been manually calculated (no idea how or to what accuracy).
81796 - unknown projection. In absence of adequate edge marking, SW and NE corners have been manually calculated (no idea how or to what accuracy).

81616 - unknown projection; minor correction to chart corners
82564 - slight adjustment made to corner calibrations
83039 - slight adjustment made to corner calibrations
83392 - slight adjustment made to corner calibrations

81127 - unknown projection; corrected datum and chart corners
81427 - wrongly specified as Mercator - corrected
83010 - corrected projection to "unknown" and multi-pixel correction to corner cals

[z.stadler]

Originally Posted by recri

Happy --

So if you have a chart that really isn't on a WGS datum, there are two possibilities:

1) You find a datum note, enter the datum or the specific correction, and after the correction the grid lines on the chart won't come out aligned to opencpn's grid lines, they'll all be displaced, along with everything else, by the correction.

2) You don't find a datum note, no correction is applied, and the grid lines on the chart will be aligned with the opencpn grid lines, but all the fixed features on the chart will be displaced by the inverse of the missing correction. So the check is to overlay onto a known WGS chart and compare the two.

So my current checking doesn't even catch case 2) and can't catch it until I have layers of charts over given features and an easy way to compare the alignment of different layers.

If the chart reads:

Datum Note

Adjustments for plotting positions obtained from satellite navigation systems based on the World Geodetic System (WGS-1972) cannot be determined for this chart

I expect the Datum should be set to "unknown".

After trying to save in various ways, it turned out that the only acceptable "Adjustments to Chart Datum" setting is to mark "no DATUM NOTE given on chart"

Maybe the text can be changed to "Chart contains no DATUM NOTE or no corrections are given"

Ze'ev

[HappySeagull]

Ze'ev..which chart?

[z.stadler]

Originally Posted by HappySeagull

Ze'ev..which chart?

71400, 71410 and a few others in that range

[recri]

Originally Posted by paulr

Region 8 checked as follows:

Thanks, Paul, I've moved them all to the censored list as they still show up as crooked.

[AussieGeoff]

Ok, I just tuned into this thread.

Where are the Region 7 original charts being sourced? I notice that a huge amount are missing from the FTP server you refer to.

For instance 75142 and it's neighbours that cover much of
South Australia?

Somewhere I can obtain those? I'd be interested in tinkering with them myself, I have played with Oziexplorer and am no stranger to GPS and georeferencing scanned paper maps and charts.

AussieGeoff

[HappySeagull]

Originally Posted by z.stadler

71400, 71410 and a few others in that range

Thanks,Ze'ev,
I see what you mean..it'd be better if "unknown" didn't bring up another dialogue box.
I haven't had a lot like these to do so it hasn't been annoying...being "Compilations" the NGA stitched from a bunch of charts from various nations together, I guess that explains why a datum was difficult.
I hope that (eventually )these "unknowns" can be co-related to the charts overlaying them and possibly corrected but maybe it never will be possible given they are a quilt of other charts....

[HappySeagull]

Originally Posted by AussieGeoff

Ok, I just tuned into this thread.

Where are the Region 7 original charts being sourced? I notice that a huge amount are missing from the FTP server you refer to.

For instance 75142 and it's neighbours that cover much of
South Australia?

Somewhere I can obtain those? I'd be interested in tinkering with them myself, I have played with Oziexplorer and am no stranger to GPS and georeferencing scanned paper maps and charts.

AussieGeoff

Geoff,we don't have 75142.We Only have what the NGA has online.We are calibrating these online as well ,using this genius tool....
NGA Charts Status Region 7 at OpenCPN.info (English)
you will need sign on to join the fun.