Georeferencing - Setting up a map from scratch
So that the modern data (imagery, elevation, GPS) fit properly
MT: Michael Posted: 10 April 2008, 11:57 AM
Lactic Turkey has made a street-park map of 2/3 of Waitakere City for the Eco City Challenge. Auckland OC and NorthWest OC no doubt have semi-topo maps of parts of Auckland too. It would be good to hear that you are all using the same mapping conventions, have all related your maps to the same national grid (NZMG or NZTM) and put your paper coordinate origin in the same place and used the same grid-magnetic angle. So that you can join your bits with OCAD-Import and know they will fit EXACTLY. Like I'm sure you've had to do in Woodhill…
MT: runningbeast Posted: 10 April 2008, 3:56 PM
Mapping conventions? national grid and paper co-ordinate origin. What mumbo jumbo is this?
We just drew the map making it symbols etc as we went using an aerial photo overlaid with council parks, contours etc as the template.
Mapping conventions? huh!
It looks good though!
MT: Bryan Posted: 13 November 2008, 3:18 PM
Don't know whether this has come up before.
I'm creating a map using New Zealand Map Grid GIS contours so the OCAD map is in NZMG coordinates (and currently oriented to Grid North).
With some clubs creating a master OCAD file combining all their maps in one file, I'm wondering how they are handling magnetic north. (for all Orienteering maps magnetic north should be at the top of the map).
If the file is oriented to magnetic north, have the clubs thought of how to handle the magnetic north variance? (eg magnetic north in Wellington is approximately 22.5 degrees east of grid north increasing at the rate of approximately 0.5 degrees every 30 years. A simple rotation of the map every few years will not work as some symbols get skewed.
Has anyone set and tested the 'Angle' value in the 'Options - Scales' Menu and what value do you use? Has anyone testing changing this value and printing maps oriented to magnetic north with symbols correctly aligned?
MT: Greg Posted: 13 November 2008, 5:23 PM
Is this really a problem? 0.5 degree over 30 years, I mean no one noticed when a nationals map was 22.5 degrees out a few years ago
MT: The Map Guy Posted: 13 November 2008, 11:51 PM
I tend to agree with Greg. Most maps are out of date after a couple of years. I thought you'd have access to NZTM co-ordinates Bryan.
I have used the “Angle” value mentioned above when importing a topo map in as a template - used the OCAD projected grid for alignment of the template. I did this to produce a magnetically oriented map. I just used the G-M angle.
MT: Bryan Posted: 14 November 2008, 7:36 AM
Yes, its not much of a problem. However, I've been mapping for 26 years (I'm showing my age) and this year I updated a map I first fieldworked in 1983 (Riverside) - it has not changed much. Also, when creating a GIS or database of maps its quite likely that reference data (like 1:50000 topographical maps and 20m contours will not change over time). I have had to rotate some maps before but not because of magnetic variation.
At my work, I have access to NZMG, NZGD2000, NZGC49, NZTM, raw lat and long, local meridional circuits and Australia is even worse. And don't mention the Chatham Islands which is like the Bermuda Triangle and a real pain as it is outside the NZMG system which most clients want (so we just place points in the middle of the sea). I aslo have access to a Canadian software package called FME which can convert between any system.
The contours I'm using are free and in NZMG and given to me in a shape file. And I'm only using Ocad 8 which is limited on coordinate systems.
MT: Michael Posted: 14 November 2008, 1:05 PM
I'm not sure if I have understood you correctly Bryan, I would have thought you were an expert! Excuse me if the following is sucking eggs. There seems to be two parts to this: setting up a map from scratch, and changing it if you decide it needs changing.
From scratch, the relationship (coords of the paper map origin, and the angle) would be first set in the scales dialogue. Then when you import contours in NZMG coordinates from shape files or DXF they are rotated appropriately so that paper north is magnetic north.
As to what that angle should be, I had an interesting time finding the authority on this. It isn't Land Information NZ, it's Geological and Nuclear Sciences. See http://www.gns.cri.nz/research/geomagnetism/ If you read this carefully, you'll realise that it isn't the magnetic-true angle that you're after, but the magnetic-grid north angle. There are graphs of its variation for NZMG (and I'm hoping that its not very different for NZTM). Somewhere between 23 and 23.5 for Wellington, maybe closer to 24 for Wairarapa.
For practical purposes I don't think a degree out is a problem. (22 degrees WAS a problem for me on the first leg of a sprint race in a complex park, though:-)) So I wouldn't be worried about the change through time. If you want to change from zero to 20-something however, I think you CAN, and there's an option whether to rotate symbols oriented to paper north or not. There is a snag with any templates you have though, the “fit” doesn't follow the rotation and they will have to be re-adjusted.
I have also changed the offset values - from their NZMG values to the NZTM values. I was prepared for complications, but it isn't the map that is moving, its the real world. (I think this makes sense in the OCAD paper-based view of things!) I'm still learning here, be good to know if I'm off beam.
MT: Bryan Posted: 14 November 2008, 2:12 PM
Thanks for that Micheal - I did go to the same web page as you but used the wrong chart (which said 22.5 for magnetic declination vs true north - instead of the chart for NZMG)
One day I might migrate to Ocad 9 which will allow me to use NZTM.
On the Ocad site there is a flash demo showing how to georeference an existing map: http://www.ocad.com/howtos/45.htm
MT: Michael Posted: 14 November 2008, 3:30 PM
I think you can probably use NZTM in version 8. My understanding is that the “real-world coords” is whatever system you want it to be. If you put into the scale dialog the NZTM coords of a point in the middle of your map, then you could import files that are referenced to NZTM. I can't prove that tho, I haven't got any NZTM DXFs.
However I suggest you do move to version 9. And right now, if you upgrade 8 to 9 (CHF380/Euro245) you'll get version 10 for free when it comes out next year. See http://ocad.com/OCAD10/OCAD_10_whats_new_en.pdf
MT: Selwyn Posted: 15 November 2008, 5:01 PM
NZTM is in OCAD8. By the way OCAD10 Beta is now released, full version early next year. There seems to some variance in magnetic variation even if you stand still! I understand that True North and Grid north are slightly different, possibly one to 1.5 degrees. and there might be a slight difference bewteen NZMG and NZTM. In Auckland using the old TOPO paper map variation with alleged increases, I calculated a variation of just under 21 deg. for 2010 The international web site gave me something about a degree less. My Garmin GPS says 19 degrees difference between NZTM and Magnetic. It's hard to a straight answer from offialdom, I presume because there are different norths. My solution has been to make all my Auckland maps 20 deg from whatever base map I'm using and all the base data is now aligned to NZTM. A single degree out here and there is not a problem for orienteers. But that one degree might cause some minor distortions in various backgrounds that are imported or if importing another OCAD map that didn't use the same deviation.
The BigMap that AOC has of Woodhill covers from Muriwai to Rimmer Rd being all the forest that our club is currently using and not using thanks to MTB and 4WD. That map is 6.6 MB in OCAD, compresses to about half that. Most mapping members use broadband and have reasonable computers. For course setters we just create a partial map for them to use. Ideally all the peripheral stuff like legend, logo and outlines are kept as a separate map. It can be imported, or the actual map used as a background to the outlay map. The 3 maps of LIDAR contours only that cover the same area of Woodhill together are 66 MB. But as the dxf files they were supplied were 400 MB. The 1m contours use only points to create “curves”. We bought 3 corresponding georeferenced TIFF orthophotos that were 260 MB each. So I am using big background maps and recently upgraded to new computer with a gutsy CAD style video card.
MT: Michael Posted: 16 November 2008, 3:36 PM
It took me a while to suss out these coordinate systems and its still hard. The important thing to note is that NZMG and NZTM are flat approximations of the earth's surface, and grid north is only the same as true north on longitude 173. If we're using grid data, its the magnetic-grid angle we should use. Seems from GNS it was 20 to 20.5 in Woodhill in 2005 so 20 is plenty close enough.
I haven't been able to think of any good reasons for finding a really accurate value, or for chasing the tiny change over time. The “real world coords” setting (including angle) will control the translation into “OCAD paper coordinates” and if the OCAD north is not exactly magnetic north its no problem. I think it may be possible to switch the real-world offsets from say NZMG to their NZTM values if it became necessary to import data on the other grid, but the angle should stay the same.
Selwyn points out that importing another OCAD map would escape the translation process, so its supposed magnetic north needs to be queried. (And in my experience so does its supposed scale!) If it has a “real-world” setting, then OCAD 9.x can import based on real-world coords, which would take care of the translation. And this is the key factor in all this - the NZ grids are known, stable coordinate systems, whereas most of our OCAD maps have no known location or orientation.