How to georeference scanned image in QGIS and vectorize items from OpenStreetMap-part 3.

Welcome in the next part of the Georeferencing post!

In the last parts we obtained scan of the shotpoints maps from australian seismic project and then we succesfully downloaded data from OpenStreetMaps to use them in georeferencing of our map.

How to georeference scanned image in QGIS and vectorize items from OpenStreetMap-part 1.

How to georeference scanned image in QGIS and vectorize items from OpenStreetMap-part 2.

Georeferencing which we performed in part2. used method of comparing two similar features both on scanned image and on map in already set up coordinate system, and then running the transformation using such input. This method had some advantages:

  • it is quick to use when you are good observer,
  • it is good to use when you have very characteristic features available like bays, complex but characteristic shoreline, extensive river network, available road network etc.
  • it is useful when you do not have any coordinates or graticules on your map (for example you have scan of simple schematic map from newspaper or some sketch)

On the other hand it also has following disadvantages:

  • it strongly depends on the GIS analyst observation abilities,
  • it can be not precise when you have two data sources in different scales (details level),
  • it requires regular GCP (ground control points) grid to keep georeferenced image not deformed

In this part we will see how we can georeference (register) our scanned map using alternate method. This method is useful when you have available such things as:

  • coordinates of the corners of your map sheet,
  • graticules or grid of geographic and geodetic coordinates,
  • known coordinates of some characteristic points on the ground.

Although this way requires some data input this method is really useful when you are not able to compare features in the ‘real world’ to the ones on your scan. For example you have map of the offshore project where are no islands or shoreline. You can only use graticules to georeference your map.

PE714546__((_[G206]_Shotpoint_Base_Map_))_

On our map you can notice that we have available coordinate grids and graticules. Now we will use them to georeference our map.

Once again open your project with Lake Colac shapefile and with previously georeferenced shotpoints map.

Then open Georeferencer (Raster->Georeferencer) from the top menu.

Capture32

Add the scanned map image clicking ‘+’ button in the top left of the screen. It will load your scanned image but also Ground Control Points from your previous image registering.

Remove them clicking Delete Point button and selecting each of them.

Capture33

When you will remove all of them click Add Button and start adding new GCP’s to your image.

Capture34.JPG

This time try to locate your coordinate grid intersections. You can read intersections coordinates from the map frame.

Capture35.JPG

This time when the registration window will open do not click ‘From map canvas’. Instead of it you can fill the X and Y with the coordinates noticed from the map itself (X=700.000, Y= 5.730.000).

Capture36.JPG

Repeat this procedure for whole the map area to obtain more than 3 points and to keep the point grid reasonably regular. I did it this way. You can register all the intersections or just the four corners of the map. More points will guarantee better control and precision.

Capture37.JPG

Notice that while you are adding the points on your map table below map fills with rows related to each point.

Now when you finished run the transformation using the same transformation parameters as we used before (output coordinate system should be set up to EPSG 20354, Helmert transformation type, Nearest neighbour resampling method). Remember to change your output raster name. I added word GRID to its name because it was registered using grid on a map.

Capture38.JPG

Now press the play button and transformation should begin. When finished georeferenced raster should appear in your map canvas (make sure that you selected ‘Load in QGIS when done’ before running transformation).

Capture39.JPG

Now probably all what you can see is only freshly georeferenced map. It is on the top of the layers tree and it covers all the layers below (shapefile with Lake Colac and previously registered scan). Let’s play with this layer properties to make other layers visible and to compare our results. To do this right click on your layer in the layers tree.

Now when you will click the arrow button from the Transparency menu marked in the picture you can select which particular color from the raster image should be transparent. It is very useful option for example when you have to join many scanned maps and their frames overlay each other. In our case we want the white color to be transparent. Click this button and select white color from the map canvas.

Capture40.JPG

You will see that the table with color code and transparency level appeared in the window. When you will click Apply you should see layers which are below. Now they are uncovered.

Now let’s change the lines colors from black to red to differentiate between new and old raster. Go to Style menu. There you can adjust colors to raster values (this was black and white reaster so the values have black colors).

Capture41.JPG

Click the first Value row and change it from black to red.

Capture42.JPG

Now you should see all three layers: green polygon which is Colac shapefile (vector file) and two raster files which we georeferenced (old with black lines and new one which is red).

Let’s use the GEarthView plugin to see how the comparison looks like in the Google Earth. Before I did this I turned off the Colac lake layer (in the layers tree with the X button) and changed the display properties of first raster layer (changed lines from black to green the same as we did above changing from black to red).

Capture43.JPG

Now what we can observe is that there are little differences between the layers registered using two georeferencing methods. The first one where we used shapes comparison seems to be more precise.

In fact it is more precise when we use it for imaging the small objects and when we are focused on particular map features. When we would like to make only Lake Colac map probably it would be better to georeference the raster using lake Colac shoreline shape. But when we would like to register map showing Lake Colac and its surroundings in smaller scale probably it would make more sense to use coordinates which we can read from map graticules.

It can’t be said which method is better. It strongly depends on the situation, data availability and your preferences. However you should always remember about advantages and disadvantages of both of the methods.

I hope you liked Georeferencing subject as this is very powerful tool to process your data and also we might use it next time!

Meanwhile do not hesitate to read other posts on practical usage of QGIS:

How to make Geological Map and Cross sections using Quantum GIS

How to make vintage map

Advertisements

One thought on “How to georeference scanned image in QGIS and vectorize items from OpenStreetMap-part 3.

Add yours

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Create a website or blog at WordPress.com

Up ↑

%d bloggers like this: