Opening and aligning .ocd maps

[dg0yt]

gandym01 reported on Sourceforge [tickets:#373]:

Today, I want to write about importing and georeferencing OCAD maps in OOM. After playing with this program for nearly a year, today is the first time that I have managed to properly georeference an .ocd map by opening it directly in OOM (as opposed to importing it from the file menu or via the template window). This took me forever to work out and so I wanted to canvas some ideas on how this process might be made better.

To help with my discussion, I have outlined the steps that I use to open an .ocd file in OOM.

• Open OOM then select File >> Open…
• Choose .ocd file then click OK.
• Select Map >> Georeferencing…
• In this window, the horizontal and vertical offsets from OCAD have been correctly imported as UTM coordinates and the declination value has been set to the "OCAD angle". Because the local coordinate system is in use, true north = grid north and therefore declination = grivation.
• Before touching anything, I write down the grivation value.
• Next, I ensure On CRS changes, keep: Projected coordinates is selected.
• For the Coordinate reference system option, I select UTM and enter the UTM zone number.
• Upon changing the CRS from local to UTM, the newly calculated grid convergence is added to the grivation. The declination remains unchanged.
• I now alter the value in the declination field (so that the grivation is set back to whatever value it was before changing the CRS).
• Upon clicking OK, I receive a message asking if I want to rotate the map content. I Click No.
• I now have a properly georefenced map.

Why is the default option to keep geographic instead of projected coordinates? When importing ocad files, the non-default option is always the one you need to choose. Could the default be changed?

In a previous response Kai said that, "Given an OCD map which is correctly aligned to magnetic north I could investigate what kind of error actually happens with that map." By correctly aligned, I take it he meant that map north is equal to magnetic north in the .ocd file. Unlike OOM, in OCAD9 there is no concept of true north or magnetic north, and therefore no grid convergence or declination. There is only grid north, map north and the angle between them, which I call the "OCAD angle", but this is actually the grivation. So, I take it that a "correctly aligned .ocd map" is one where the "OCAD angle" is set equal to the grivation value?

The problem here is that when an .ocd map is opened (via File >> Open...), the OCAD angle value is improperly assigned to the declination field. However, since the CRS is not set, it is impossible at this stage to calculate the grid convergence, and therefore the true declination is simply incalculable. When the CRS is specified by the user, the incorrect declination value is preserved and a new grivation value is calculated by summing the declination and the grid convergence values. This is not what should happen. Rather, the declination should be calculated by subtracting the newly calculated grid convergence from the "OCAD angle".

If I modify the declination value to achieve the right grivation by hand, then this triggers a question that I should not need to answer to achieve the overall goal. The question is, "The declination has been changed. Do you want to rotate the map accordingly, too?" The correct answer is No, but not everyone is going to guess this answer correctly first time.

So, what can be done? Firstly, it appears to me that (almost) all of the information required to do the job should be available from the .ocd itself without any user involvement. You should not need to specify that you want to retain the projected coordinates, nor set the declination by hand, nor should you have to answer the question about rotating the map (which shouldn't arise in the first place). You shouldn't always need to specify the CRS, but, perhaps the current importer is not sophisticated enough to interpret/convert the embedded CRS information? Also, if the CRS is not specified in the .ocd file then the user would need to be prompted and given an opportunity to provide this information.

So, at most, the user could be queried for the CRS, but all other information should be obtainable from the .ocd file. My suggestion therefore is that, upon opening an .ocd file (via File >> Open...), a dialogue is called. This dialogue could query the user to optionally provide an appropriate CRS (or perhaps ask the user to confirm a CRS obtained from the .ocd file). With this information, the grid convergence could be calculated, allowing the correct declination value to be determined and set. Also, the question of rotating the map does not need to be asked or answered.

I hope these suggestions are useful and not just something obvious that hasn't been looked at yet. It would be great if something like this made it into the next version.

gandym

[dg0yt]

• Milestone: 0.5.93 --> 0.5.94

[dg0yt]

• status: new --> assigned
• assigned_to: Kai Pastor

[dg0yt]

The OCD grid rotation was explicitly used as grivation. But the dialog did not properly deal with the "initially unkwown" state of the declination. This is changed now.

The georeferencing dialog needs testing in map and template georeferencing when 0.5.94 is released.

This ticket remains open for dealing with CRS specification / GPS corodinates from OCD files.

#333 is closely related and provides test data.

[dg0yt]

gandym01 posted on Sourceforge:

Great. Thanks. I look forward to seeing how this was implemented and testing it out.

gandym

[dg0yt]

• Milestone: 0.5.94 --> 0.6.1

[dg0yt]

Part of this issue may be fixed as of #784.

[dg0yt]

This ticket remains open for dealing with CRS specification / GPS corodinates from OCD files.

This should be resolved in the unstable builds.