How to georeference a map in UTM / WGS 84 (geographic) to WGS84 / UTM?

I have scanned UTM WGS 84 toposheets, with a geographic grid. I want to georeference them in WGS 84 / UTM (therefore with UTM coordinates). Should I:

1. Calculate the UTM coordinates from the geographic coordinates and enter them for the GCPs, or
2. Enter the geographic coordinates for the GCPs and then reproject in UTM?

I think "1" is cumbersome but more logical as there is no back and forth change from planar to geographic projection.

• Could you clarify how will you calculate UTM for Toposheet since you mention they are with geographic grids. Feb 1, 2015 at 9:11
• If you know the geographic coordinates, you can easily look up the right UTM zone for that part of the world. Feb 1, 2015 at 10:07
• @AndreJ, Sorry if i fail to clear myself enough. I am not concerned with UTM zone but the reading of GCPs in UTM zone(since you have only geographic grids labelled in toposheets) to georeference sheets against these readings. Feb 1, 2015 at 11:11

Here is an alternative which you can do using QGIS:

2. Set the project CRS to WGS84
3. Create a vector grid with the same spacing as your toposheet uses
4. Densify the vector grid
5. Change the project CRS to the UTM zone for the area of interest

1. Start the georeferencer and load the scanned map
2. Georeference the imprinted grid vertices with those from your self-made grid

This way you use the degree information from the map, compared to real-world degrees, but in the desired UTM projection. So this is your option 1), but let the computer do the calculating...

• Thanks, very useful. But I still don't understand how geographic coordinates can on a UTM map without (visible) distortion at 1/50,000. Feb 1, 2015 at 11:06
• On my example, you see that the degree grid is rotated against the UTM background. That's the distortion. And one degree of longitude is smaller (in UTM meters) than one degree latitude (unless you are near the equator). I created the grid using WGS84 without any distortion, then changed the project CRS to UTM to let QGIS do the reprojection on-the-fly. Feb 1, 2015 at 14:48
• @Oulala the closer you are from the equator, the more UTM lat and long will be similar in UTM. So if you are in the South of India, you will merely see a difference Feb 1, 2015 at 20:24

IMHO the best option is 1.

From your description, the coordinate system of your topo sheet is in UTM (a projected coordinate system is necessary to represent data on a flat surface). On top of that you might have an WGS 84 coordinate system that is used for the graticule (with units in degree because it is a geographic coordinate system). This coordinate frame does not necessarily match the projection of the map, as it is probaly your case.

So, the geometry of your map is in UTM and your GCP's are in WGS 84. In order to avoid distortions of your map, you should convert yours GCP coordinates to UTM and then you can make an affine transform (3 GCP needed + one for RMS estimate). There are many sites to do this, here is one.

This is a guess based on my understanding of your description, but I am pretty confident. To be 100% sure, please add the full text describing your reference system and a print screen of a part of your graticule.

• Apart from the distortion, your resulting image may look weired when you reproject to geographic and back to UTM. Jan 31, 2015 at 17:52
• @AndreJ, Yes my case is USA is in the ASIA (in degree readings) when i add grids-graticules in the map. Jan 31, 2015 at 18:34
• @msi_g Then you have not specified the right UTM zone. Feb 1, 2015 at 7:25
• @AndreJ Thnaks,It is at LEGEND, I mean i have same PCS reading for different GCS. Feb 1, 2015 at 7:50
• @radouxju If I got OP, how he/she will get exact PCS to enter against GCPs since he/she has geographic grids only (in the sheet)-could you clarify a bit. Feb 1, 2015 at 7:54

I think you should go for your option 2(since you mentioned you have only geographical grids in toposheets). UTM is a projectd coordinates system(2d plane) which comes into effective when geographic coordinate system(3d space) is defined beforehand. IMHO,Almost all works(e.g. reading from GPS etc) direction is firstly project in geographic then onto projected if needed. So the notion of distortion due to back and forth transformation is not that much of to be considerable.

See features(point) with same Projected Coordinate Systems Reading reside in different Geographic Location

Just to visualize(roughly) above two points on WGS Datum--

N.B. In fact there are same projected coordinate reading for different UTM (geographic) zones. for example if i take toposheets' geographic (as OP says with a geographic grid. ) coordinate system and get projected one for this geographic (say X,Y=1034973.46806954, 2458590.2185667) and georeference against this reading it will not be correct because, for a single projected coordinate system (say my 1034973.46806954, 2458590.2185667) there are different geographic coordinate systems (e.g. UTM Zone 23,24,25,26,27,28 etc) then if you go with projected coordinate system firstly then how it will be correctedly georeferenced (in a shortcut it will give incorrect degrees-minute-second, i mean real USA is in ASIA in map).

• Although UTM is a worldwide system, you have to specify which UTM zone you want to use. The UTM coordinates will only be valuable for this zone. You can not reproject a worldwide map to a single UTM zone. Feb 1, 2015 at 7:24
• Yap, World feature class is in WGS84 Only.UTM uses WGS84 as GCS, My points are UTM 20 and 46 and also uses WGS84.Care reprojecting image with geographic grids.In fact i mean there exists same PCS reading for Different GCS and that's all. Feb 1, 2015 at 7:56
• Thanks @msi_g. I would calculate the UTM coordinates, then assign the zone and normally it would land in the right place. But the core of the question (beginner) is how geographic coordinates can be put on a UTM projection. There should be distortion (scale is 1/50000). Feb 1, 2015 at 11:01