Georeferencing rasters
Sometimes, you have a paper map, an image of a map from the Internet, or even a raster file with projection data included. When working with these types of data, the first thing you'll need to do is reference them to existing spatial data so that they will work with your other data and GIS tools. This recipe will walk you through the process to reference your raster (image) data, called georeferencing.
Getting ready
You'll need a raster that lacks spatial reference information; that is, unknown projection according to QGIS. You'll also need a second layer (reference map) that is known and you can use for reference points. The exception to this is, if you have a paper map that has coordinates marked on it or a spatial dataset that just didn't come with a reference file but you happen to know its CRS/SRS definition. Load your reference map in QGIS.
This book's data includes a scanned USGS topographic map that's missing its o38121e7.tif projection information. This map is from Davis, CA, so the example data has plenty of other possible reference layers you could use, for example, the streets would be a good choice.
How to do it…
On the Raster menu, open the Georeferencing tool and perform the following steps:
- Use the file dialog to open your unknown map in the Georeferencing tool.
- Create a Ground Control Point (GCP) of matches between your start coordinates and end coordinates.
- Add a point in your unknown map with GCP Add +. You can now enter the coordinates (that is, if it's a paper map with known coordinates marked on it), or you can select a match from the main QGIS window reference layer.
- Repeat this process to find at least four matches. If you want to get a really good fit do between 10-20 matches.
- (Optional) Save your GCPs as a text file for future reference and troubleshooting:
- Now, choose Transformation Settings, as follows:
- You have a choice here. Generally, you'll want to use Polynomial. If you set 4+ points for the first order, 6+ points for the second order and 10+ points for the third order, The second order is the currently recommend one. This will be discussed in the There's more… section of this recipe.
- Set Target SRS to the same projection as the reference layer. (In this case, this is EPSG:26910 UTM Zone 10n)
- Output Raster should be a different name from the original so that you can easily identify it.
- When you're happy with your list of GCPs click on Start Georeferencing in File or on the green triangular button.
How it works…
A mathematical function is created based on the differences between your two sets of points. This function is then applied to the whole image, stretching it in an attempt to fit. This is basically a translation or projection from one coordinate system to another.
There's more…
Picking transformation types can be a little tricky, the list in QGIS is currently in alphabetical order and not the recommended order. Polynomial 2 and Thin-plate-spline (TPS) are probably the two most common choices. Polynomial 1 is great when you just have minor shift, zooming (scale), and rotation. When you have old well-made maps in consistent projections, this will apply the least amount of change. Polynomial 2 picks up from here and handles consistent distortion. Both of these provide you with an error estimate as the Residual or RMSE (Root Mean Square Error). TPS handles variable distortion, varying it's correction around each control point. This will almost always result in the best fit, at least through the GCPs that you provide. However, because it varies at every GCP location, you can't calculate an error estimate and it may actually overfit (create new distortion). TPS is best for hand-drawn maps, nonflat scans of maps, or other variable distorted sources. Polynomial methods are good for sources that had high accuracy and reference marks to begin with.
If you really want a good match, once you have all your points, check the RMSE values in the table at the bottom. Generally, you want this near or less than 1. If you have a point with a huge value, consider deleting it or redoing it. You can move existing points, and a line will be drawn in the direction of the estimated error. So, go back over the high values, zoom in extra close, and use the GCP move option.
Sometimes, just changing your transformation type will help, as shown in the following screenshot that compares Polynomial 1 versus Polynomial 2 for the same set of GCP:
Polynomial 1
Note the residual values difference when changing to Polynomial 2 (assuming that you have the minimum number of points to use Polynomial 2):
Polynomial 2
Tip
Resampling methods can also have a big impact on how the output looks. Some of the methods are more aggressive about trying to smooth out distortions. If you're not sure, stick with the default nearest neighbor. This will copy the value of the nearest pixel from the original to a new square pixel in the output.
See also
- When performing georeferencing in a setting where you need it to be very accurate (science and surveying), you should read up on the different transformations and what RMSE values are good for your type of data. Refer to the general GIS or Remote Sensing textbooks for more information.
- For full details of all the features of the QGIS georeferencer, refer to the online manual at http://docs.qgis.org/2.8/en/docs/user_manual/plugins/plugins_georeferencer.html.
- The QGIS documentation has some basic information about how to pick transformation type at http://docs.qgis.org/2.8/en/docs/user_manual/plugins/plugins_georeferencer.html#available-transformation-algorithms.