Conversion from ANG (Australian National Grid)

[snowygreybeard]

Does anyone know of free software/calculator for converting coordinates from ANG to something more contemporary? ANG is Australian National Grid, used on old inch:mile maps for example. I found a $$$ app, but hard to justify the dollars for the amount I need.

[peregrinator]

https://binaryearth.net/AusDatumTool/

[snowygreybeard]

Thanks Peregrinator. Its not free, but not too much. The free trial might be all I need.
Unfortunately the free online version doesn't support ANG.

[Allchin09]

What are you trying to do?

Precise transformation of coordinates from those old military maps to current day is hard due to the inaccuracies of the original surveys.

This means there isn't a global uniform formula you can use to transform from that datum to a current one.

I've spent a bunch of time trying to create better methods and I've had some recent success with the help of AI to OCR and clean up old survey records which I can then process and come up with the amount of transformation needed at various points throughout the state.

To be honest, the easiest way is to read it off one of the old maps which has been georeferenced.

If you use www.bushwalkingmaps.com, you can see most of the old military maps with that grid reference system.
I have also added an approximate 'ANG' grid based on my most recent calculations which you can also use. It doesn't always line up with the maps (sometimes the maps aren't georeferenced precisely enough)
Eg, Osborne College is approx at 236 443 ANG.

I have some older ramblings here on the topic, it's not updated with some of my latest work - https://maps.ozultimate.com/wiki/georef

[snowygreybeard]

Thanks for that info Allchin09. I clearly have lots to learn in this field. Yet to get my head around the vagaries of ANG. I naively thought there would be a straight formula conversion.
My immediate desire is to georef old SMA maps of Snowy Mts, then on QGIS trace old tracks, huts and camps - to be put onto my GPS for reference in the field. I have a bunch of A3 sized 4":mile maps that often don't have well identified points for georeferencing - other than the actual grid.

[andrewp]

You could try this tool to convert from ANG.

http://www.binaryearth.net/AusDatumTool/index.php

[Allchin09]

My immediate desire is to georef old SMA maps of Snowy Mts, then on QGIS trace old tracks, huts and camps - to be put onto my GPS for reference in the field. I have a bunch of A3 sized 4":mile maps that often don't have well identified points for georeferencing - other than the actual grid.

Yes, we'd all be happier if there was a formula!

Hmmm, if you want any real accuracy for your tracings, you probably can't easily georeference based on the grid alone, it would be better to try and do it off the map features if possible. Sounds like that might be tricky though.

I've worked with Anthony Dunk, the creator of the Australian Datum Tool, on this problem in the past. His built in ANG datum only gives approximate coordinates and is computed based on converting coordinates in a specific area.
You can see some of his notes here https://www.binaryearth.net/AusDatumToo ... lGrid.html

One option you have is to try calculate the transformation parameters by using his ComputeDatumShift tool, say based on trig points from the 1:63,360 inch to the mile maps and the current day trig coordinates for your area of interest and then use those in QGIS to define a proj transformation / coordinate reference system which you can try use for georeferencing.
https://www.binaryearth.net/ComputeDatumShift/

[snowygreybeard]

As it turns out I've led myself up the garden path. The SMA 1inch map series do not appear to use ANG after all. I'll look into following the option you suggest. Thanks.

[Allchin09]

Must be AGD66 (particularly I've it's dated after 1966). If so, you can georeference it correctly in QGIS by setting your source coordinate reference system to AGD66/AMG66 and entering in the coordinates off the map.

[snowygreybeard]

One such map was compiled in 1960. Grid is 1 mile. Origin of grid is 300 miles West and 300 miles South of intersection of 148 deg 30 min E and 34 deg 0 min S. Transverse Mercator projection. Any ideas anyone?

[Allchin09]

Very odd. The ANG (or whatever is used on the inch to mile military maps) use yards for the grid and offset of origin, not miles.

Doing some googling, I think you might be talking about these maps?
https://nla.gov.au/nla.obj-652961107/view
Seem to align to your description and have the same origin. And yes, it all seems to be in miles, not yards, which is very odd!

I'd just suggest georeferencing in QGIS off 6 or so common points using a basic transformation like the Helmert or Projective transformation type.

I might have to think about how you could define some transformation using the grid reference, but your issue is that lat/lons prior to 1966 still weren't as well defined as they are now. You could assume that they align with 1966 but they could be off a bit in that area. One way to try would be to georeference just off the lat lon references on the map itself (they align with the black and white sections of the border) and assume them to be AGD66.

[snowygreybeard]

Thanks for that. I actually have two types of SMA maps. One is the series example you gave above (1":mile). Unfortunately they are on paper and I don't have access to A1 (or A0?) scanner. Was hoping to scan A4 sections of interest. Doing this and using lat/long as you suggest is possible, but messy when in the middle of the sheet. (EDIT: Doh, realised the digital copy is available via your Trove link. OK, something to do tonight)
My other series are 2":mile, A3 sized - which I can scan. Although some have one lat/long point, they don't have the markings to give me the required 3 or 4 points.(Edit: Realised they are actually 4":mile maps, but copied half size!)